Brendan Nyhan posted yesterday about his article in the just-released special issue of The Forum on the politics of health care reform. There are several compelling articles in the issue by notable scholars, including Representative David Price.
My own contribution to the issue (along with co-author David Eckles) is an expansion of an earlier post on this blog. Here is the key take-away from our piece:
While a large majority of Americans did see rising health care costs as a problem, very few of these same people thought that reform would improve this situation, and when it came to whether people supported or opposed the reform plan, it was the anticipated costs of the legislation, not concerns about current rising costs, that appeared most salient to Americans. Ultimately, Democrats passed health care reform legislation in spite of their inability to secure significant public support for the plan. Yet their efforts to mitigate the effects of loss aversion on public support for the proposal may have kept even more Americans from opposing the legislation, and if Republicans mount a serious attempt to repeal the reform law, it will be Democrats who are appealing to the public's
aversion to risk and loss.
For the most part, the public agreed that rising health care costs were a major issue and that something had to be done to curtail those costs. However, they also tended to agree with Republicans that the health care reform legislation was not going to help limit those costs. In fact, a significant proportion thought it was going to make them worse. And, as the figure below indicates, prospective views about how the legislation would influence costs had a much more influential role in structuring opinion on the health care reform legislation than did concerns about current rising costs.
Another major point of our article is the importance of loss aversion; that is, the public's tendency to over-value what they already have and under-value what they do not yet own. This tendency worked against Democratic efforts to win public support for health care reform, but it is also why we argue that now that people have been given health care reform, it will likely be quite difficult for Republicans to attempt to take it away.
Check out our piece here, and all of the other great contributions here. (Free registration is required for access).
When I moved from DC to Amherst in August I was looking forward to the charm of a small New England college town and the relative affordability of housing (compared to prices inside the beltway, at least). But what I knew I'd miss the most was living at the center of the political universe. Well, for one day at least, I get to re-live the excitement as all eyes turn to the Bay State.
Warning: what follows is entirely un-scientific and is, accordingly, of little use to understand what might happen when the polls close tonight.
I've been surprised over the past few weeks at how much of a ground game Brown seems to have in Western Massachusetts compared to Coakely. Last weekend, I had to drive to nearby Belchertown for a swim meet and we passed countless yard signs for Brown as well as a small rally of Brown supporters. Not a single sign for Coakley. Today I had to travel to Boston for an appointment and I was 40 minutes into the drive before I saw my first Coakley sign.
Of course, once I got into Boston things changed quite a bit. Coakley signs were much more prominent and the handful of polling sites I passed were packed. This is consistent with what the news has been reporting regarding high turnout. But what was most interesting from my vantage point (and the vantage point of any pollster trying to determine who will and will not vote today) was a conversation I had with a young Democratic store clerk who, upon finding out that I was a political science professor, started complaining about how he ended up having to get up early today to go vote when he had been planning all along to skip this election. I've had similar conversations with a number of Democrats over the past few days; people who had no intention of voting a week ago, but now feel compelled to do so. These people are not at all excited about the Coakley campaign, but they suddenly feel as though they have to go out to vote in an election they were planning on skipping.
A friend sent me a couple of links earlier pointing to pundits and pollsters who are taking last night's results as evidence for the merits of IVR polling. First off, as Mark noted earlier, it is a bit too early to be making such comparisons. With regard to the claims being made about IVR polling in particular, I would add the following points:
First, there is no way to control for other reasons that these polls might have generated different results, including different approaches to screening for likely voters and how undecideds are dealt with. With regard to the latter issue, it is important to note that the pollsters using live interviewing in New Jersey were showing more than twice the percentage of undecideds as those using IVR.
This leads to a second important point (related to the first): comparing these pollsters based on the final result presupposes that each pollster that has been entered into this fictitious competition was actually trying to get the final result correct in the first place. If that was the goal, then it seems as though each polling firm would have allocated all of their undecided respondents into one camp or another.
Third, one of the reasons for concerns with IVR polling is that citizens with only a cell phone cannot be reached by these pollsters and these citizens now comprise at least one-fifth of the population. Yet, while the cell-only problem may generally be an issue for IVR technology (and for live interview pollsters who aren't calling cell phones), it is less of a problem for polling on elections, and particularly in low turnout elections. This is because the types of people that do not have landlines are less likely to be voters (and particularly less likely to be voting in low turnout elections). Ultimately, an off-year low turnout election may actually be less of a challenge for IVR-based polls because the non-coverage bias should be smaller for these contests. Where these polls may run into greater challenges is when they attempt to make inferences about the American public rather than registered (or likely) voters.
The New Yorker has an interesting piece on how the public's aversion to losses (or loss aversion) limits the extent to which they are willing to favor health care reform. That piece and some others that preceded it are worth reading to understand one reason that Americans may support the general idea of reforming the health care system, but then express far less support when confronted with the possibility that their own health care plans may be affected. The bottom line is that individuals tend to value what they already have to a much greater extent than what they might gain (this is often called the endowment effect). This means, for example, that people are far less willing to part with an item that they already have than they are to forgo receiving that same item if it has not yet been in their possession.
"In the health care debate, loss aversion helps to color the public's perception of potential reform. A recent Gallup poll found a clear majority of Americans favor health care reform in the coming year. But when pressed on specific aspects of the health care, Americans are decidedly loss averse. Almost 90 percent of Americans want to be able to choose any doctor or hospital they like, and 77 percent of Americans say it is important to have the option to keep the health insurance plan they have now. In sum, we may want change and reform - but not at the cost of any of our current options."
In other words, those that already have health insurance probably overvalue that insurance relative to what would be available to them under health care reform legislation, and this may be driving down support for reform.
Fortunately, the survey released by the Economist yesterday provides a nice addendum to these readings by illustrating how loss aversion can significantly alter public opinion depending on how a question is framed. In this survey, the sample was split randomly into halves. The first half of the sample was asked to choose which of the following plans they preferred:
"A plan with no lifetime limit on beneﬁts."
"A plan that limited the total amount of beneﬁts in your lifetime to $1 million, but saved you $1000 per year."
Four out of five respondents (80%) answering the question framed in this way selected the first option. They'd much rather have a plan with no limit on benefits than save $1,000, but be subjected to a $1 million lifetime limit.
The second half of the sample chose between these options:
"A plan that limited the total amount of beneﬁts in your lifetime to $1 million."
"A plan with no lifetime limit on beneﬁts, but cost you an additional $1000 per year."
Functionally, these options are equivalent to those presented to the first half of the sample. In the first presentation, the limited plan will increase the respondent's wealth by $1000 per year by saving him or her that money; in the second presentation, the limited plan will increase the respondent's wealth by $1000 per year because that respondent will not have to pay the cost of the unlimited plan. However, the different framing of the options (emphasizing "savings" rather than "cost") is critical. Among those choosing from the second set of options opinion was more closely divided--44% chose the plan with limited benefits while 56% chose the unlimited benefits option. In short, more Americans wanted the unlimited plan when it meant forgoing a savings of $1,000 per year than when it meant incurring a cost of $1,000 per year.
It is also important to note that the changes in how the options are framed do not affect all groups equally. In particular, loss aversion appears to be conditioned by income. This makes sense since wealthier respondents may not be as sensitive to a $1,000 per year change in their wealth as those with lower incomes. To demonstrate the relationship, the chart below compares the percentage of respondents who would choose a plan with no lifetime limit depending on whether they received the question with the "savings" frame or the "cost" frame. Respondents are broken down into three income categories.
What stands out from this chart is that respondents in each income category are much more likely to chose the option with no lifetime limit when they received the question with the savings frame. However, under the "cost" frame, responses differed more significantly across income categories. Thus, among respondents making less than $40,000, support for the plan with no lifetime limit was 32 percentage points higher when that plan was presented as a way to forgo a savings of $1,000 rather than incurring a cost of $1,000. The framing effects were much smaller for those with higher incomes.
Of course, these aren't actually the choices being presented to Americans during the health care reform debate, but this survey experiment does provide a neat way of illustrating not only how the framing of health care reform as a potential loss can affect support for the measure, but also among which groups those frames will be most effective. Indeed, the New Yorker article ends by noting that it may still be possible to gain public support for health care reform despite the public's tendency toward loss aversion:
"The key may be to work with, rather than against, people's desire for security. That's surely one reason that Obama has consistently promised people that if they like the health insurance they currently have they can keep it. This promise will make whatever reform we get more inefficient and less comprehensive, but it also assuages people's anxieties. It might even be possible to use the endowment effect and the status-quo bias in the argument for change. After all, although people tend to feel that they own their health insurance, their entitlement is distinctly tenuous...Changing the system so that individuals can get affordable health care, while banning bad behavior on the part of insurance companies, will actually make it more likely, not less, that people will get to preserve their current level of coverage."
For the public to support health care reform, the reform needs to be framed as something that will help keep most individuals (who do have insurance) from losing what they already have. Furthermore, the analysis of the Economist survey suggests that individuals with lower incomes are most likely to respond to such an attempt to re-frame the debate in this way. This is notable since there is much ground to be gained among these individuals. In fact, the same survey shows that respondents in the lowest income group are substantially more likely than others to be unsure about whether the health care reform plan would make them better or worse off. This group appears to have their minds least made up on health care reform and their opinions may be the most susceptible to the efforts by both sides to frame this issue during the coming weeks and months.
Political pundits generally settle on a shared view of a campaign, one that includes a story about which groups each candidate worked hardest to win votes from. But how does the general public perceive the candidates' campaign strategies?
In 2008, I included a battery on the Cooperative Congressional Election Study which asked 1,000 American adults to indicate which types of people each candidate had focused more attention on (the actual wording of the question was "During the presidential election campaign, which of the following groups do you think [Barack Obama/John McCain] has focused most of his attention on?") Respondents could select up to five groups from a list of 21 and each respondent was asked to complete this exercise for both candidates.
I am using these data for a project I'm working on looking at targeting during campaigns; however, I thought that I'd share some initial results here. The chart below aggregates the responses to these questions to show how the public viewed both candidates' campaigns. The chart shows what proportion of the public that thought that Obama (on the y-axis) or McCain (on the x-axis) had focused on winning the votes of each group.
Groups in the upper left hand corner are those that a large proportion of the public thought the Obama campaign focused on targeting, but only a small share thought McCain did. Clustered far up in that corner are young adults, lower income Americans, and African Americans. Also near that top left corner are liberals. None of these groups are surprising to see in this corner, though they may be there for different reasons.
Groups in the bottom right corner are those that a large share of voters thought McCain targeted but which fewer thought Obama focused on. These groups included whites, conservatives, and upper income Americans. No surprises here either.
In the top right corner are groups that Americans thought both candidates focused on winning votes from. Interestingly, there are very few groups in this area, with middle income Americans standing mostly alone. Aside from this group, the public did not appear to identify too many groups that they thought both candidates were trying to win over.
Some other interesting findings from this chart:
More Americans thought that McCain tried to win the votes of women than Obama. It is interesting to ponder how big a role the Pallin selection was in affecting this perception. It is also worth noting that while more Americans thought that McCain was trying to win the votes of women, the gender gap strongly favored Obama in the actual voting.
Women are not the only group where the public's view of the candidates' strategies didn't quite match with the actual success of the candidate among that group. For example, Obama edged out McCain among Americans earning $150,000 or more. He also won big among those describing themselves as ideological moderates.
It is also interesting to see where Born-Again Christians fall on this chart. Despite the publicity Obama gained for targeting young evangelicals, few in the public actually credited him with trying to win over the votes of this group. However, perhaps more intriguing is the question of where this group might have fallen along the x-axis in the 2004 election. While between 20 and 30% of Americans thought McCain, it seems likely that this number is significantly lower than it would have been for Bush in '04.
Finally, there appears to be some polarization in these perceptions. In particular, note how far apart the income and racial groups are distributed on this chart. The public viewed African Americans as being almost the exclusive domain of the Obama campaign while whites were overwhelmingly viewed as being only targeted by McCain. Furthermore, the Obama campaign was viewed as being the only campaign focusing on lower income Americans while McCain was the only candidate viewed as focusing on those with higher incomes. On the other hand, both candidates were viewed as targeting middle income Americans.
These perceptions undoubtedly vary depending on whether a respondent is or isn't a part of each particular group. For example, those with higher incomes may have been more likely to think that Obama was focusing more attention on those with lower incomes compared to those who actually have lower incomes. I'll be exploring these dynamics when I analyze the data in more detail.
Overall, the chart provides some interesting insight into how the public viewed the candidates' strategies and raises an interesting question...how comparable would these perceptions be to the judgments of journalists and political pundits who follow the campaign for a living?
One of the major themes in the paper is that understanding the cell-only population is about more than just age. In fact, residential mobility has a strong influence on whether someone has shed their landline. Even after controlling for age and a litany of other demographic variables, we find that respondents who moved within the last year were 24 percentage points more likely to be cell only than those who had lived in the same residence for at least five years. Renters, singles, and those without children were also much more likely to be cell-only.
Our explanation for this pattern:
"There are several reasons that highly mobile Americans may be more likely to go without landlines. First, whenever someone moves from one residence to another, they have an opportunity to reassess their phone needs. Thus, the act of moving provides an opportunity for individuals to shed their landlines. Second, mobile Americans may choose a CPO lifestyle because cell phone numbers tend to be more portable than landlines. When moving from one metropolitan area to another, individuals must change their landline phone number, but do not need to change their cell number. This may provide an incentive for choosing not to maintain a landline in a new residence. Third, those with fewer family and community ties may feel less of a need to have multiple phone lines on which they can be reached by members of their social networks. "
The fact that the cell-only public tends to be more mobile has some important political consequences. Some highlights:
The difference in the percentage of landline and cell-only respondents who reported being registered was fairly small--over 95% in both groups. However, there was a much larger gap in actual registration rates (66.8% versus 53.9%). Since cell-onlys are more likely to have moved recently, they may not have successfully registered to vote at their new addresses despite the fact that they may think they are registered.
Cell-only respondents were significantly more likely to have problems with their registration when attempting to vote. In 2008, over 7% of cell-only respondents indicated that there was a problem with their registration when they attempted to vote, compared to fewer than 4% of respondents with landlines.
Cell-only respondents were more than twice as likely as those with landlines to report that neither campaign contacted them. In short, this group is much less likely to be subjected to mobilization efforts from the campaigns.
Cell-onlys are politically distinct on a variety of measures. However, this distinctiveness is somewhat muted when demographic controls are taken into account. Interestingly, the largest differences between cell-only and landline respondents are not on issues or ideological self-placement, but on reported vote choices.
Ultimately, we argue that weighting for standard demographic measures such as age, education, income, and race may not be sufficient. Pollsters relying on landline samples may want to consider weighting by other factors such as time in residency, renter/home owner, and marital status. But check out the full paper for a more detailed discussion of all of these points.
The survey research community is focusing intently on the challenges posed by the fast-growing share of Americans who are cell-phone-onlys (CPOs). In fact, there are 40 papers being presented on the topic at the AAPOR conference next month. One of the practical issues faced by pollsters is whether the cost of reaching CPOs is worth the payoff. Last week, Scott Keeter, Mike Dimock, and Leah Christian hosted a forum at Pew during which they discussed this tradeoff. But pollsters aren't the only people who have to make cost-benefit decisions when it comes to deciding whether to attempt to contact CPOs. Campaign organizations must make the same calculation.
So how well did the campaigns do at contacting CPOs during the 2008 campaign? The chart below compares the percentage of those with landlines and cell-onlys who reported being contacted by a campaign representative in 2008. The data comes from the National Election Study (NES), which uses residential sampling and face-to-face interviews to interview both landline and CPO respondents. In the chart below, the blue bars show the percentage of each group that reported being contacted while the black lines represent 95% confidence intervals for these percentages.
The chart shows that CPOs were much less likely to be contacted by the campaigns than people with landlines. Over half of landline respondents reported being contacted compared to less than one-in-three CPOs. This sizable difference holds up even when controlling for age, income, education, partisanship, and a variety of other factors.
The next chart (below) indicates that for those CPOs who were contacted, the contact tended to come overwhelmingly from Democrats. Over 80% of CPOs who were reached by the campaigns were contacted by the Democratic side while just a little over one-third were reached by Republicans. Republicans were significantly more competitive with Democrats when it came to contacting those with landlines.
Unfortunately, the NES did not include questions asking respondents how they were contacted by the campaigns. But a subset of respondents to the 2008 Cooperative Congressional Election Study (which I've analyzed in previousposts) were asked these questions. The chart below plots the responses for those who had landlines compared to CPOs.
CPOs who were contacted by one of the campaigns were significantly less likely to have had that contact over the phone compared to those with landlines. Otherwise, there were not major differences between how landline and CPO respondents were contacted. CPOs were somewhat more likely to get an email while those with landlines were a bit more likely to receive snail mail, but neither of these differences are large. The percentage being contacted in-person or by text message were nearly identical for both groups.
Overall, the findings from these surveys suggest that shedding your landline may help you avoid those pesky campaign calls in future election years. While Democrats were a little more successful than Republicans in reaching CPOs, the cell-only crowd was almost as successful avoiding campaign volunteers as they were hiding from pollsters.
Steve Ansolabehere and I have been working over the past few weeks on a paper we are writing for the AAPOR conference next month. Over the next couple of weeks, I'll share some of our preliminary findings here and I wanted to lead off today by presenting some comparative data we put together from three different surveys with distinct approaches to sampling.
The National Election Study, which has been around since 1948, is a labor- and time-intensive survey that uses a residential sampling approach and conducts face-to-face interviews. The Cooperative Congressional Election Study is an web-based survey conducted by YouGov/Polimetrix. The respondents opt-in to the study and Polimetrix uses a sample matching methodology where they first select a random target sample and then attempt to find a match for each respondent in their pool of opt-in respondents. Finally, Pew's survey work is fairly well known to regular Pollster.com readers. Their efforts to incorporate CPOs into their samples has relied on a dual-frame approach where they randomly select landlines and then cellular lines to build their sample.
So what happens when we compare these surveys, each of which takes a different sampling approach? Each survey I'll present data on here was conducted during the presidential campaign last year. For starters, the chart below compares each survey's estimate of the size of the cell phone only population after sample weights are applied. Note that Pew includes phone status as one of their weighting criteria, but the NES and CCES do not.
CPOs make up 19.7% of the CCES sample after weighting; 17.9% of the Pew sample and 17.3% of the NES sample are CPOs (again, after weights are applied). Thus, the differences are relatively small. Also note that the 95% confidence intervals for each estimate (represented by the darker lines) overlap.
So each of the surveys provides a similar estimate of the size of the CPO population, but what about the composition of that group? The table below compares landline and CPO respondents on a wide array of demographic and socioeconomic factors.
There is a lot of information in the table, but the different surveys are fairly similar across most measures. A few differences do stand out, however. After weighting, the CPO respondents reached by Pew and CCES were less likely to have children, less likely to be married, and less likely to be home-owners than those reached by the NES. Pew's CPOs also tended to have lower incomes and were somewhat more likely to be racial/ethnic minorities. Perhaps NES's face-to-face approach is more likely to pick up CPOs who have "settled down" relative to CCES or Pew.
Speaking of "settling down," one other item deserves attention; that is the information on residential mobility at the bottom of the table. While we weren't able to find this information in any of Pew's surveys conducted in October or November, the data were available for the CCES and NES. Not surprisingly, residential mobility is strongly related to whether one is a CPO or not. Even controlling for age, people who have moved recently are much more likely to have shed their landline and gone with just their cell. In fact, this relationship holds up when you control for all of the other demographic variables in the table. This is something I'll post more about later, but you can probably imagine that there are some significant political consequences arising from the fact that CPOs tend to move around much more frequently.
Anyone who has ever watched Deal or No Deal has noticed that some people are far more willing to take risks than others. Not only does a person's tolerance for risk affect their decisions about whether to open another suitcase on a game show, but it also influences countless daily decisions like what to eat or whether to drive over the speed limit. But what might a person's tolerance for risk have to do with their political views?
This past Fall, I was part of a team that had a module on the 2008 Cooperative Congressional Election Study. To shed some light on the role that risk might have on political preferences, we borrowed a set of questions developed by economists to gauge an individual's tolerance for risk. Essentially, the questions ask individuals the circumstances under which they would be willing to take a new (and equally good) job. The chart below shows the distribution of responses.
Not surprisingly, the results suggest that the American public is quite risk averse. Most Americans were not willing to take a new job even if the potential for increasing their income was greater than the potential income loss. In fact, over half of the respondents would not take a job even if it offered them an even chance of either doubling their income or cutting it by just 20%. These findings are similar to those of other studies that have looked at how risk tolerant (or intolerant) the public is. But what does this have to do with politics? Take a look at the charts below which show the partisan/presidential vote breakdown among each of the four levels of risk tolerance.
How a person feels about risk is related to that person's choice of party as well as their vote choice. In particular, as one becomes less tolerant of risk, they become more likely to affiliate with the Republican Party (and less likely to be a Democrat). The relationship was even stronger when it came to vote choice. The most risk tolerant respondents preferred Obama by more than a 2-to-1 margin. The most risk averse respondents went for McCain by a margin of 6%.This relationship persisted even when I controlled for other factors that tend to influence vote choice.
When you think about it, there are lots of reasons that an individual's tolerance for risk would influence their political views. It makes sense that the most risk averse Americans were less likely than others to get behind the candidate who was viewed as relatively inexperienced. (Risk averse Americans are probably more likely to vote for incumbents as well.) It also seems logical that those who are more averse to risk would be more likely to affiliate with the party that tends to be more conservative.
There are several recentacademicpapers on this topic (including my own). Yet, despite the growing evidence showing that risk matters for how people think about politics, pollsters rarely include questions that allow us to capture respondents' feelings about risk. Is it time that they started to do so?
This post is part of Pollster.com's week-long series on Stan Greenberg's new book,Dispatches from the War Room.
I've enjoyed reading through Greenberg's thought-provoking book over the last several days. The exercise has led me to think not just about the relationship between pollsters and their clients, but also about the role of pollsters and political consultants in our democratic system. Greenberg makes an interesting case that pollsters play an integral role in helping to link elected officials to those they govern, and his argument is convincing.
In his lead-off post on Tuesday, Mark referred to V.O. Key's famous definition of public opinion. The quote was familiar for me (and many others, I'm sure) as I use it as a point of departure for my PhD course on political behavior. Key described public opinion as "those opinions held by private citizens which governments find it prudent to heed." I've added the emphasis to the second part of the sentence because I think it is crucial. For Key, "opinions held by private citizens" do not, on their own, rise to the level of public opinion. Rather, only those opinions that capture the attention of political elites can be classified as public opinion. But which attitudes are these? Or, perhaps more to the point, whose attitudes are these?
Greenberg points out in his book that he initially gained notoriety by reporting on his study of white union members in Michigan (p. 19-20). The report caused a "storm" and led to a lot of focus in Democratic circles about how to win back the so-called "Reagan Democrats." The DLC hired Greenberg to gain more insight into how to win over this group and much of Clinton's appeal was his ability to speak to these working class whites. "Reagan Democrats" were suddenly exercising significant influence over how Democrats would campaign and, ultimately, how they would govern.
But time and resources are finite and politics is a zero-sum game. When pollsters draw politicians' attention to groups like "Reagan Democrats," other groups will necessarily get less attention and, therefore, may be less represented in government. Consider this quote from a story by Politico's Avi Zenilman just after the verdict in our most recent election:
"Unions, Hispanic groups, the Netroots, progressive organizing coalitions, single women, working women, youth, the religious left -- to name just a few -- all claim to have played a vital role in electing Barack Obama. And each says he owes them for that role."
These groups are actively lobbying for elected officials to notice their importance in the last election because they understand that being viewed in that light will lead to more influence over government. But it strikes me that pollsters may have a special role in telling their clients which of these groups have the most valid claims and should be paid special attention to.
This brings me back to how influential the role of a pollster can be during the time between elections. At the end of the book, Greenberg notes that he was always careful to limit his role to reporting on public opinion and avoided devising policy prescriptions (as Clinton himself said, "he wouldn't tell me what to do"). This suggests a relatively innocuous role for a pollster--one that is merely informational. However, if pollsters play a key role in identifying particular groups--"swing voters" or otherwise--that politicians should be particularly concerned with, then doesn't their role become far more influential? After all, the voters that politicians think they need to appeal to will likely have a major influence on the types of policies those politicians promote. By choosing to shine the spotlight on some groups, don't pollsters (and ultimately their clients) leave other groups in the dark?
So I'm interested in hearing the extent to which Greenberg thinks that his role as a pollster was influential simply by identifying some groups as "swing voters" but not others. There are so many different ways to slice and dice the public to identify the next influential group (whether it be "Reagan Democrats" or "married mums"); does Greenberg see any groups who were largely overlooked by government because they never rose to "swing voter" status? And if he was going to go back out into the country now, as he did after the 1984 election, and focus attention on a particular group of citizens, which group does he think would be most deserving of his attention?
Frequent Pollster.com readers will know that the Cell Phone Only (CPO) population is a subject I blogged onfrequentlyduringthecampaign. One of the reasons for this interest is because there is still a lot we don't know about CPOs at this point (I like a good mystery). In this post, I want to present a little analysis that demonstrates just how much we still have to learn about CPOs.
CPOs wouldn't be nearly as fascinating if their distinctiveness could be easily explained by their age or other basic demographic or political factors. One view is that CPOs are just younger, more urban, and more mobile and that once you account for these factors, CPOs behave pretty much just like their landline counterparts. Last year, however, a Pew report suggested that the differences might not be that simple when they discovered that weighting alone might not be enough to account for CPOs.
To explain why this is the case, I used some of the data Pew relied on to publish that report (in this case, Pew's June 2008 Voter Attitudes Survey). In this survey, Pew interviewed respondents both on landline phones and on cell phones (see more on the methodology here). What I wanted to know was whether CPOs are really that different from landline users once you account for demographic and political factors?
To answer this question, I estimated a multivariate statistical model that would allow me to take into account any demographic or political information that the Pew survey collected that I could imagine would explain the difference between CPOs and the rest of the public. I controlled for age, gender, education, ethnicity, race, income, home ownership, marital status, and whether the respondent lived in an urban area. In addition to these demographic factors, I also accounted for the partisan affiliation of the respondent. If these factors explained the difference between CPOs and landline respondents, then we wouldn't expect to find any measurable difference between these groups once we've controlled for them.
If you enjoy reading output from statistical models, you can get that information View image. However, the key information is presented in the chart below. The chart shows the probability of a landline or CPO respondent registering a vote preference for Obama, McCain, or neither candidate after controlling for the demographic and political factors.
This chart indicates that even after controlling for all of the demographic and political factors listed above, CPOs still had distinctive vote preferences relative to those with landlines in their homes. While landline respondents were 49% likely to prefer Obama in June, CPOs were 65% likely to do so. These differences can't be explained as a result of CPOs being younger, or because they were single, or because they lived in urban areas, or even because they were more likely to be Democrats. Essentially, if you had two people who were the same on all of the factors mentioned above, the one without a landline would still be more likely to support Obama than the one with a landline.
If it isn't age, income, education, or even mobility, then what makes CPOs distinctive from those with landlines? Is it something more inherent about embracing a CPO lifestyle? Perhaps it is an outlook on life that makes CPOs more willing to cast off traditions and venture into something new? Perhaps a higher tolerance for taking risks and embracing change? Or something else entirely? We still don't have a good handle on what makes CPOs so distinctive, which is why this makes for such a great mystery.
(Note: I'm using this Pew data to examine another reason why CPOs may cause pollsters so much trouble--because it may be more difficult to pin down whether they are actually going to vote. I'll present some analysis on that topic in my next post.)
Back in December, a SurveyUSA polldrew some attention and fueled the hype regarding the large crowds expected for Tuesday's inauguration ceremonies. In that poll, an incredible 12% of respondents reported that they were planning to "attend the inauguration of President Obama in Washington, DC." That would translate into over a half million Atlantans making the 500+ mile flight (600+ mile drive) to DC for the ceremony.
Of course, most of us probably know people who originally said that they were going to attend but later backed out for any number of reasons. But just during the past week, SurveyUSA asked adults in the New York and Los Angeles media markets whether they planned "to go to Washington for the inauguration." Remarkably, 9% of New Yorkers reported that they were planning such a trip while 6% of those in Los Angeles said the same. For some perspective, the adult population of the New York media market is nearly 16 million while the Los Angeles market has nearly over 12 million adults. Thus, if we took these polls at face value, DC would be expecting over 1.4 million visitors from New York and over 700,000 from Los Angeles. Are 2.1 million coming to the district from just these two markets? Not likely.
One problem with a question like this one may be that it lends itself to social desirability bias. As we know, citizens tend to over-report the extent to which they will (or did) vote in elections. In a similar way, some respondents may be proclaiming that they will attend the inauguration when they don't have any real intention of going. They may do so because they hear of so many others who are attending and they feel as though it is something they should be doing as well.
However, even if citizens gave honest answers to the question about whether they were going to attend the inauguration, surveys still wouldn't be a very useful instrument for generating crowd estimates. For instance, let's assume a national poll showed that 2% of American adults were attending the inauguration ceremonies. That would translate into approximately 4 million adults. However, even if you had a margin of error of just 2%, you would only be able to confidently say that the expected crowd may be anywhere between a few thousand (which would be the equivalent of a calm summer day in DC) to as many as 8 million (a number that would cripple the area's transportation infrastructure). Thus, polls such as these may provide a fun way of capturing excitement about the event, but they are not very useful for estimating how large the crowd will actually be.
Set aside for a moment the fact that President-Elect Obama's support for a college football playoff is the right position*, it is also good politics. Obama has brought up his support for a playoff on several occasions, and there are two good reasons for him to continue to do so--his position on the issue is supported by a large majority of college football fans, and college football fans are made up of groups that Obama underperformed with in the presidential election.
In the most recent poll I could find on the topic, Gallup recorded overwhelming support for a playoff among college football fans. In this January 2007 survey, 69% of college football fans supported a playoff that would involve at least four teams while an additional 16% favored a one-game playoff among the top two teams after the bowl games. Thus, only 15% of college football fans want to keep the present BCS system.
As the chart below indicates, support for a college football playoff is nothing new; college football fans have consistently supported such a system at least over the past two decades. In 1994, 72% of college football fans expressed a preference for a playoff system. And despite the fact that the BCS was designed to reduce such clamoring, support for a playoff remains high (and is increasing). Thus, Obama's position on this issue puts him in-step with well over over two-thirds of college football fans; politically, announcing his support for a playoff is as difficult a decision as coming out in favor of ice cream and sunny days.
Not only is Obama's position on the side of the vast majority of college football fans, but these fans also tend to be the kinds of voters that Obama would like to make inroads with in 2012. According to Gallup, College football is America's third most popular sport, with 53% of Americans reporting that they are college football fans. College football fans tend to be married men, and they are more common in the South and Midwest than in other parts of the country. As the chart below indicates, these are precisely the types of groups that Obama under-performed with in the 2008 presidential election.
Of course, aside from Pete Carroll,Mack Brown, and Kyle Whittingham, there aren't very many people out there who are going to be evaluating the Obama presidency based on this issue. But Obama's public statements have sparked a lot of conversation about a playoff on ESPN and other sporting outlets (in addition to the regular news media) and it doesn't hurt to have his name frequently linked to a position that is this popular, particularly among groups who don't exactly make up his base. So it won't be surprising if we keep hearing Obama speaking up for a playoff whenever he has a chance.
Throughout the campaign, much was made of the tremendous ground organization that Obama had built. Yet, according to the exit polls, Obama's organization did not contact a higher percentage of voters than Kerry's did in 2004. In both 2004 and 2008, voters were asked "Did anyone call you or talk to you in person on behalf of either major presidential campaign about coming out to vote?" In 2008, 13% said that the Obama campaign had contacted them while 13% reported that both campaigns had done so. That means that 26% of voters nationwide had been contacted by the Obama campaign. This figure is the same as Kerry's contact rate among 2004 voters. In fact, nationally, the major difference between 2004 and 2008 was that the Republican contact rate dropped. In 2004, 24% of voters reported that they were contacted by the Bush campaign but in 2008, just 19% were contacted by McCain.
Of course, national figures may mask more significant patterns at the state-level. I was able to get statewide results for the question on campaign contact in 17 states. These results are plotted in the chart below for every state where the data were available. Nearly all of these were states that Obama ultimately won; the only exceptions are West Virginia and Missouri.
Every state except West Virginia falls above the diagonal line, indicating that the Obama campaign contacted more voters than McCain in 16 of these 17 states. However, the margin of Obama's advantage in campaign contact did vary. In several states, Obama held an advantage similar to that captured by the national contact margin (7%). However, Obama had at least a 12% advantage in the percentage of voters contacted in Iowa, Virginia, Indiana, Colorado, and Nevada. Of course, Iowa and Nevada were states where the Obama campaign had spent significant time and effort organizing for the January caucus events; this effort appeared to pay off in the general election as 50% of Nevada voters and 42% of those in Iowa reported having been contacted by the Obama campaign. Virginia and Indiana were states where Obama invested heavily during the primary and general election campaign. 50% of Virginia voters and 37% of Indiana voters reported contact from the Obama campaign. Finally, the Democratic convention was held in Colorado, which may have helped make that the state where Obama contacted more voters than anywhere else--51% of Coloradans reported being contacted by the Obama campaign.
With the exception of New Mexico, reported contact rates are available for each of the states that Obama flipped from red to blue in 2008. The chart below compares Obama's advantage in contacting voters in each of the states he flipped with his national advantage on this measure.
Note that of the three states that come closest to Obama's national advantage, two are states that have been targeted by both parties in recent presidential elections. Thus, it may be the case that Republicans already had the infrastructure to be more competitive with Obama's ground game in traditional battleground states like Florida and Ohio, but they were unable to catch up with the organization Obama built in new newer battleground states.
Finally, it is difficult to trace the precise impact of this ground game effort relative to other factors during the campaign. But for fun, the chart below plots the Obama contact advantage in each of the 17 states against the change in the Democratic margin in that state from 2004 to 2008.
Again, we shouldn't draw too many conclusions from this plot, but it does appear that Obama tended to improve more on Kerry's 2004 electoral performance in states where he held a larger advantage in voter contact.
Ultimately, the state-level data provide important context about Obama's ground game advantage. Obama dominated the ground game in most of the states he turned from red to blue in 2008, particularly the newer battleground states like Nevada, Colorado, and Indiana. This organizational advantage was undoubtedly one of the major factors behind such a large vote swing in those states.
Now that most of the national vote has been tabulated, we can get a pretty good sense of which pollsters came closest to pegging the final popular vote. As Mark mentioned in an earlier post, several others have donethisalready, but I thought I'd create these plots for Pollster.com readers.
The final national poll results from individual pollsters are plotted below (these are the last 19 national polls listed on Pollster.com's national trend page). The pollsters represented with red dots are those that included cell phone only (CPO) respondents in their sample. The Obama vote is represented by the y-axis and the McCain vote is the x-axis (UPDATE: I've updated the plots to reflect the updated vote share of 52.5% for Obama and 46.2% for McCain). The horizontal red line is the actual vote that Obama garnered while the vertical line indicates McCain's share of the vote. The closer a poll is located to where the two red lines meet, the more accurate that poll was in predicting the final outcome.
Note that every pre-election poll plotted here underestimated McCain's support. However, the big winners wereRasmussen and Pew, both of whom estimated a 52-46% advantage for Obama. The Pew poll included CPO respondents while the Rasmussen survey did not. CNN and Ipsos/McClatchy also came quite close by estimating a 53-46% advantage for Obama (UPDATE: These polls now come just as close as the Pew/Rasmussen polls). Neither survey reached the CPO population. Indeed, the plot reveals no clear pattern with regard to the CPO issue. Polls including CPO respondents did not appear to be any more accurate than those only reaching landlines.
Before the election, I separated out the Pollster.com national trend into surveys including the CPO population and those who were only calling landlines. The plot below looks at how each of these trends performed.
The trend based on surveys including the CPO population did slightly better at estimating Obama's vote but worse at gauging McCain's support. Overall, the CPO trend was slightly further off the mark than the landline trend.
Of course, there are any number of other factors at play with these different surveys (such as different likely voter screens, weighting, etc), so we can't draw any definitive conclusions from this analysis. But there is no obvious pattern from these initial results that indicate that including CPO respondents helped improve polling accuracy.
Update: Updated to reflect changes in popular vote.
The New York Times has a great tool for comparing vote patterns among a few basic demographic groups going back to 1980. Here are a few patterns that stand out:
Women: Obama won women by 13% over McCain. That is the second largest gender gap since 1980 (only Clinton's 16% advantage in 1996 was larger).
Men: The national exit poll currently shows Obama with a 1% edge over McCain among men. The only other Democrat to win men since 1980 was Bill Clinton in 1992. But that was in a three-candidate race, so Obama's share of the male vote (49%) was the highest for any Democratic candidate during the last eight presidential elections.
African Americans: African Americans made up a larger share (13%) of the electorate than they had in any of the past eight presidential elections. In addition, Obama won 96% of the black vote whereas previous Democratic nominees failed to do better than 90%.
Hispanics: Hispanic voters continued a pattern that began in 2006 by renewing their strong support for Democratic candidates. Obama captured two-thirds of Hispanic voters, who made up 8% of the electorate.
Young Voters: Obama won two-thirds of the 18-29 age group. This is the fifth straight presidential election that Democrats have won the youth vote. However, Obama's margin with this group is substantially greater than any previous Democratic nominee. In fact, no candidate has won any of the standard exit poll age groups by as big a margin as Obama won the 18-29 vote in this election.
Mark is back and will be starting a new thread now.
8:32 - [Brian] Early national exit polls indicate that Obama has won 68% of the Hispanic vote. In 2004, the exit polls showed Kerry winning just 54%.
8:36 - [Brian] The national exit polls are also suggesting an uptick in African American turnout. In 2004, African Americans comprised 11% of the electorate. That figure is up to 13% in the current national exit polls. Another point: Kerry won 88% of the black vote; Obama is winning 96% of that vote.
8:44 - [Brian] Note the exit poll margin in Mississippi: showing just a 1% edge for McCain. African Americans make up 36% of the electorate according to early exit polls (not much different from the 35% in 2004), but the exit polls are showing 99% of African Americans voting for Obama in that state. Just 17% of whites voting for Obama, but that is up slightly from 14% for Kerry in 2004.
8:53 - [Brian] Obama appears to be out-performing Kerry among every age group except those 65 and over. Kerry lost that group by 5%, the early national exit polls show Obama losing them by 10%.
9:00 - [Brian] Bush won the suburbs narrowly in 2004; the early national exit polls are suggesting Obama will win the suburbs in 2008. But more significant may be that in 2004, Bush won the voters in rural areas by 15%; this year, Obama is only losing rural voters by 6%.
Mark is back and will be starting a new thread now.
I think we'll have a pretty good sense of where the night is headed after the early polls close. Of course, as Mark has warned use repeatedly, we want to be careful about reading too much into early exit polls before the weighting is adjusted to account for actual turnout and results. But here are some things to look for in a few states with early poll closings.
Indiana is a state where Obama should presumably have benefited from the protracted nomination campaign and the massive organization he has build in that state. Thus, if turnout among young adults is going to increase markedly, it should be obvious here first. According to exit polls, in 2004, 14% of the Indiana electorate was between the ages of 18-29. We have to be careful with early exit poll figures since young voters may be more enthusiastic (and, thus, more likely to show up in early exit poll results), but if that figure goes up significantly in this election, then that is probably the first evidence we will have that young voters are turning out at higher rates in this election.
Virginia may tell us more about this election than any other state. Not only does the map look very difficult for McCain if he loses Virginia (particularly if the networks can call it relatively quickly), but the demographics in Virginia can provide us with some useful insight into what may happen in other states. In 2004, exit polls indicated that African Americans made up 21% of the Virginia electorate. Will that figure improve in 2008 and, if so, by how much? What will the party id figures look like? In 2004, 39% of voters said they were Republicans compared to 35% who were Democrats. Democrats would like (and probably expect) to see those numbers flip in Virginia just as they are looking for party id gains in other high growth states like North Carolina, Colorado, and Nevada.
Finally, returning to African American turnout, here is a remarkable statistic. In 2004, 834,331 African Americans voted in Georgia's presidential election. Already this year, 705,203 African Americans have voted early in that state. African Americans make up about 30% of registered voters in Georgia but in 2004 they comprised just 25% of the electorate. It appears as if we are well on our way to seeing a huge surge in African American turnout in Georgia, and when the polls close there at 7pm, we should have a pretty good sense of whether African Americans will make up 30% or more of the electorate in the state. If so, there is a reasonable chance that Obama can win Georgia and that a landslide may be in the offing. To do this, he needs to perform slightly better among whites than Kerry did. According to exit polls, Kerry won just 23% of the white vote in 2004; Obama would need 27-30% of the white vote to capitalize on the high turnout among blacks (or he would need Bob Barr to peel away a significant share of McCain's support). This is still a bit of a long shot, but Georgia has one of the first poll closings, so it will give us something to look for during the 7pm-8pm hour.
A little over a week ago, I posted two different national trend estimates: one for pollsters who were reaching cell-phone-only (CPO) respondents by calling cell phones in addition to landlines and one for those who were only calling landlines. At the time, Obama's lead in the trend of pollsters accounting for CPOs was about 3% wider than among those who were only calling landlines.
Here are the two different trend estimates as they stand on the even on eve of the election.
National Trend Estimate for Pollsters Reaching Cell Phone Only Respondents
National Trend Estimate for Pollsters not Reaching Cell phone Only Respondents
If anything, the difference between the two trends is greater now than it was a week and a half ago. As of Monday morning, Obama's lead was 4.2% larger in the national trend accounting for the CPO population than it was among the landline-only polls. In addition, while the landline-only polls are showing some late tightening in the national trend, the surveys reaching CPO respondents do not show any such tightening.
While many of the major national polling firms have made a great effort to include CPOs in their polling this fall, it is important to keep in mind that most of the state-level surveys fail to reach CPOs. Thus, there is a possibility that the state trend estimates may be under-estimating Obama's support. What happens if we try to account for the CPO effect in the statewide trends?
The three charts below show Obama's margin in the states currently classified as leaning or toss up on the Pollster.com map. The first chart shows the Obama margins according to the Pollster.com trend estimates as of Monday morning. The second chart makes a conservative CPO adjustment by adding 2% to Obama's margin in each state. And the third chart makes a 4% adjustment to the CPO to mimic the current difference we see between the two national trends.
If you make no CPO adjustment and give each state to the candidate currently leading, Obama wins 367 electoral votes, narrowly losing Indiana, Montana, and Georgia and narrowly winning North Carolina and Missouri. Making a conservative CPO adjustment by adding 2% to Obama's margin in each state pushes Indiana and Montana into Obama's column, giving him 381 electoral votes. Finally, if you make a 4% CPO adjustment to Obama's margins in each state (based on the differences in the national trends), Georgia suddenly shifts into Obama's column, giving him 396 electoral votes. Of course, it is important to keep in mind that the cell phone only population is not evenly distributed across the 50 states so not all states will be affected in the same way. But if you believe that there is a cell phone only effect that the state trends are not capturing, then states like Virginia, Nevada, and Ohio are not even that close right now and Obama has a good chance of winning in Indiana, Montana, Georgia, and possibly even Arizona.
Tomorrow night, we will have a better sense of how much a difference the CPO population has made in polling this race. Which of the national trends presented here comes closer to pegging the final popular vote tally? Does Obama win some of the states where the polls show him behind by a few percentage points? The bigger the Obama margin in the national vote and electoral college, the more likely that some pollsters missed some of his support by failing to reach the CPO population.
The McCain campaign thought they were on to something when "Joe the Plumber" confronted Obama about redistributing the wealth. Yet, there is little evidence that this argument has actually helped them make any real inroads into Obama's support. For example, the ABC News/Washington Post poll asks likely voters which candidate they trust more to handle the issue of taxes. The McCain campaign began making the "Joe the Plumber"/"spreading the wealth" argument during the third presidential debate on October 15th. Since then, there has been no significant movement on the question of which candidate the electorate trusts more on taxes. Obama maintains about a 10% edge on this issue, just as he did before the rise of "Joe the Plumber."
Why has the argument failed to gain traction? To take a stab at this question, I went back to some polling data I happen to have on my hard drive from a February 2003 NPR/Kaiser Family Foundation/Kennedy School of Government Taxes Survey. The survey is useful for addressing this question because it delved deep into philosophical issues about the U.S. tax system.
First off, let's start with what Americans know about the tax system. The survey revealed that only one of three registered voters said that they had heard the term "progressive taxes" and knew what it meant (that figure has probably increased during this campaign). However, while a majority of Americans didn't know the terminology, most (70%) did understand that people who make more are taxed at a higher rate than those with lower incomes.*
But how do voters feel about a tax system that redistributes wealth? Respondents were asked whether it was "the responsibility of government to reduce the differences in income between people with high incomes and people with low incomes." 37% of registered voters strongly disagreed that this was the government's responsibility while 28% strongly agreed (the rest were fairly evenly divided between somewhat agreeing and somewhat disagreeing). Thus, on first glance, it seems like McCain's critique of income redistribution should be a successful one. But let's drill a little deeper by breaking down responses by party identification.
While a majority of Republicans strongly disagree with the idea that the government should work to reduce income disparities, independents are far more divided on the issue. In fact, nearly as many independents strongly agree that the government should be doing this as strongly disagree. Thus, once you move beyond the Republican base, the criticizing the government's role in wealth redistribution appears to be more of a wash. And at this stage of the campaign, McCain needs to win over those independent voters to gain ground on Obama.
In fact, when the McCain campaign criticizes Obama for raising taxes (or letting tax cuts lapse, as the Obama campaign prefers to frame it) for high income Americans, they may be treading on dangerous ground. When respondents to this survey were asked whether high income Americans pay their fair share in taxes, 58% of registered voters said that they paid less than their fair share. Just 18% of registered voters said that high income Americans paid more than their fair share. As the chart below indicates, this sentiment was particularly prominent among Democrats and independents. From this perspective, it is not surprising that McCain hasn't gotten much traction by criticizing the fact that Obama wants to increase taxes for high income Americans. Most Americans, particularly those beyond the Republican base, appear to think that high income people should be shouldering more of the tax burden than they are.
Another problem with using "Joe the Plumber" to criticize the redistribution of wealth is that this argument doesn't seem to have any particular appeal for the demographic "Joe the Plumber" is supposed to represent--working class whites. The chart below shows strong agreement/disagreement for the government's role in reducing income disparities among whites making less than $75,000 per year and those making more than $150,000 per year.
It is clear from this figure that McCain's argument should be a big hit among white voters making more than $150,000 per year. Nearly 70% of this group strongly disagrees that the government should be reducing income disparities. However, among whites making less than $75,000 per year, the argument has much less resonance. In fact, these voters are just as likely to strongly support a tax system that reduces income disparities as they are to strongly oppose it. Likewise, 64% of these voters said that high income people do not pay their fair share in taxes. The problem for McCain becomes even more pronounced since there are about three times as many whites making less than $75,000 per year as making more than $150,000.
Thus, these data indicate that McCain hasn't gained much ground with the tax argument for two reasons. First, critiques of income redistribution and higher taxes for those in the top income brackets appear to mostly resonate with Republicans (who are already supporting McCain) and they have far less appeal for independents. Second, the argument also fails because the symbol doesn't fit the argument very well. Working class whites are just as likely to strongly favor the government's role in income redistribution as they are to oppose it and most among this group feel as though high income Americans aren't paying their fair share in taxes. Thus, "Joe the Plumber's" views on taxes are not really representative of the views of the demographic he is supposed to symbolize. Ultimately, the tax arguments made by the McCain campaign may resonate with his base, but they are doing little to help him make inroads into Obama's support among independents.
* As an interesting aside, nearly half of the 70% of respondents who knew that higher income Americans were taxed at a higher rate also said that middle income Americans pay the highest percentage of their income in taxes. Only 35% of those who understood that the tax system was progressive later said that high income Americans pay the highest percentage. This discrepancy is curious and may suggest that while the public understands that higher income Americans are supposed to pay a higher rate in theory, they may also believe that the upper class uses loopholes to avoid paying their fair share in practice. This would explain why "closing loopholes" is a point that Obama frequently returns to during tax debates.
On October 1st I noted that we weren't likely to see much movement in the polls in October. This was based on survey data from the past two campaigns that indicated that few voters tend to change their minds once they have settled on a candidate. Of course, if an October comeback was fairly unlikely, then a final week comeback is undoubtedly a longer shot.
According to this site's national trend estimate, Obama's margin is almost twice as large as the percentage of undecided voters left. Thus, at this point, McCain's only path to victory involves attracting support from voters who are currently planning to vote for Obama. This certainly isn't impossible. After all, pre-election polls ask voters which candidate they would vote for if the election were held today. Just because a respondent says they would vote for a particular candidate if the election were held today does not necessarily mean they have made a final decision on that candidate. For example, the ABC News/Washington Post tracking poll indicates that 9% of respondents who have chosen a candidate say that there is a chance they could change their minds. Should this give the McCain campaign any hope?
To answer this question, I looked at exit polls from the last four presidential contests. In each year, the exit polls included a question asking respondents: "When did you finally decide who to vote for in the presidential election?" This question provides a way of capturing which voters had not completely settled on their vote choice until the last week of the campaign. The chart below plots the percentage of late deciders across the past four presidential elections.
Interestingly, voters appear to be settling on their vote choices earlier in recent campaigns. In 1992, one-quarter of the electorate said that they did not come to a final decision until the final week of the campaign and the figure was about 30% in 1996. However, by 2004, nearly nine in ten voters reported that they had settled on their presidential vote choice before the final week of the campaign. Whether it is because of increasingly longer campaigns or heightened polarization, the fact that voters are making their final decisions earlier in recent contests does not bode well for the McCain campaign. In short, there are fewer late deciders to win over than there were in previous years.
Nevertheless, according to the ABC News/Washington Post survey, 8% of Obama supporters say that they could still change their minds (11% of McCain supporters say the same). Is there a chance that McCain can win over those Obama supporters that still have doubts while holding on to his own soft support? Recent history suggests that such an outcome is unlikely. The chart below breaks down the vote choices made by those who said that they came to a final decision during the last week of the campaign.
According to the evidence from the exit polls, in every election since 1992 Democrats have fared better than Republicans among late deciders. Of course, it may be the case that in these elections Democratic support was softer than Republican support going into the final week. But in none of the last four elections did late deciders break more for the Republican candidate than for the Democrat. Once again, this pattern does not give the McCain campaign much hope.
Overall, current polling and recent history suggests that there is little hope for a final week comeback. There are fewer late deciders in recent elections than there have been in the past, a pattern that seems to be holding in this contest. Furthermore, late deciders have tended to break more for the Democratic candidate in previous elections, not the Republican. Thus, the prospect of a McCain comeback seemed quite unlikely at the beginning of October and it appears to be truly improbable now.
One can get lost in the deluge of polls which, just this week, show anything from a narrow 1% Obama lead (AP-Gfk) to a substantial margin of 14% (Pew). One pattern that seems to have become particularly evident this week is that the polls showing the biggest leads for Obama tend to be those that are polling the cell phone only population (such as Pew, CBS/New York Times, and ABC/Washington Post). We know from the recent Pew report that excluding cell phone only respondents from the sampling frame reduces Obama's margin by 2-3%, even when the sample is weighted. But how does this affect the national trend estimate, which takes into account all polling?
One of the great features of the new interactive tracking charts available on this site is the ability to select or remove particular pollsters. I used this feature to create two national trend estimates--one including only pollsters that include cell phone only respondents, and one including all other pollsters.
National Trend Estimate for Pollsters Reaching Cell Phone Only Respondents
National Trend Estimate for Pollsters not Reaching Cell phone Only Respondents
The comparison between the two trends is remarkably consistent with what the Pew Report would lead us to expect. While the trend that includes pollsters not calling cell phones shows an Obama advantage in the 6-7% range, the trend for those reaching cell phone only respondents shows an Obama lead greater than 10%. Obama's support increases by almost 3% in the national trend that includes polls reaching cell phone only respondents while McCain's support decreases by about 1%.
The difference between these two trends is hardly trivial since an extra 3-4% in the national vote could very well mean that several additional states tip in Obama's favor, producing a substantial electoral college landslide (keep in mind that most statewide polls are not including cell phone only respondents). If we assume that polls that reach the cell phone only population are more accurate, then Obama's lead may very well be in double-digits. But on November 4th, it will be worth checking back on these two trends to see whether the cell phone only pollsters actually do fare better in predicting the election outcome.
UPDATE: One of the nice things about the dynamic charts is that they will continue to update themselves. Thus, if you want to keep track of the differences between the separate trends for the next 11 days, you can bookmark this post and keep checking in.
There has been endless speculation about the role that race will play, if any, in polling the 2008 presidential race. I highly recommend Mark Blumenthal's postson the issue as a starting point. In this post, I wanted to address one particular question about this dynamic suggested by the Biden quote above. That is, are the remaining undecided voters--those that will make their decisions between now and November 4th--more likely to break for McCain because of an unwillingness to come to grips with voting for a black president?
We obviously won't know the answer to this question until November 4th, but perhaps we can gain some insight from Harold Ford's unsuccessful race for the Senate two years ago. Ford lost a tight contest to Bob Corker in 2006. According to Pollster.com, the last five polls in that race showed an average 4 point lead for Corker (50-46%). The final outcome was a 51-48% win for Corker, suggesting that late deciders made little difference.
But we can consult the National Election Pool exit poll from that contest to gain a better sense of how race might have affected late deciders (those who said they made their vote choice during the last week and a half of the campaign). As the table below indicates, it was hardly the case that late deciders flocked away from the African American candidate. In fact, Ford performed better among late deciders than he did among those who had made up their minds earlier in the campaign. (Note: The exit poll showed a virtual tie despite the fact that Corker won by 3% of the vote). Evidently the late deciders were not predominantly citizens who were unable to come to grips with voting for an African American candidate.
But what about looking for race effects where they are most likely to exist? First of all, about 10% of the late deciders in Tennessee were African American voters who are unlikely to have been susceptible to concerns about Ford's race. Furthermore, particular subgroups of white citizens are more likely to be influenced by a candidate's race compared to others. Specifically, less educated, lower income, and older whites may have been particularly likely to break against Ford at the end. We might also have expected to find such a dynamic among rural whites, particularly those living in the eastern (Appalachian) part of the state. Comparisons between early and late deciders for each of these groups appear in the table below:
The differences between early deciders and late deciders are opposite of what we would expect if there was a race effect among late deciders. Whites who decided within the last week and a half of the campaign were actually 8% more likely to vote for Ford than those who made up their minds earlier. The same pattern held for less educated whites, rural whites, and whites living in eastern Tennessee. The only two groups where Ford did not do better among late deciders was for low income whites and older whites. But even in this case, Ford performed about as well as he did with early deciders, not significantly worse.
What does this mean for the presidential race? It depends on the extent to which you think the case of Tennessee in 2006 can be applied to the 2008 presidential contest. On one hand, the demography of Tennessee would seem to make it a good place to look for race effects among late deciders. On the other hand, electing someone to the Senate in a midterm election is a bit different from electing a president. But if you believe the comparison, then the experience from Tennessee in 2006 would suggest that there is little reason to expect late deciders to break against Obama because of his race. To the contrary, Ford actually did slightly better among late deciders in 2006, something that allowed him to finish a few points closer than pre-election polling had indicated. If a similar dynamic works for Obama, he may win by a larger, not smaller, margin than the current polling suggests.
"With a restive electorate, with an economy that's sort of chugging around, with a war in the background, at the end of eight years of Republican rule in the White House, Obama should be way ahead."
Karl Rove, Face the Nation, August 10th
The conventional wisdom put forth by political pundits is that Obama should be winning this election by huge margins. The argument is that with a weak economy and an unpopular president, the Democratic nominee should be "crushing" the Republican nominee and the fact that Obama hasn't had double-digit leads throughout means that there must be something wrong. But how well should the Democrat really be doing in this race? Is Obama really underachieving as much as most people assume?
To get a sense of whether Obama is underachieving, we first need to know how the Democrat should be doing in this election. This can be difficult to quantify, but one way of doing so is to borrow from political science models that are used to project presidential election outcomes. Each presidential election year, a handful of political scientists publish their predictions in PS: Political Science & Politics. Most of these models include at least some of the following: various measures of the status of the economy, an accounting for which party holds the White House (and how long they have been there), and the president's approval rating. This year, all but one model predicted a Democratic victory in the presidential race.
These models can provide us with some guidance on how well Obama should be performing in this race. After all, they are based on the same economic and political factors that pundits have used as evidence for their claims that Obama is underachieving. In the chart below, I show the percentage of the two-party vote that each model predicted the Democrat would win in this election. (I excluded 3 predictions that use polling data from earlier in the campaign as one of their predictors. Since a measure of Obama's support earlier in the race is included in the model, their predictions wouldn't provide good estimates of how the Democrat should be performing). The chart also includes the variables being used in each model to generate the predictions. The horizontal line indicates the share of the two-party support that Obama is currently winning in the Pollster.com national trend.
The predictions indicate that the Democratic nominee should win anywhere between 50.1% and 58.2% of the two-party vote. Currently, Obama is receiving 54% of the two-party support in the Pollster.com trend estimate. That places him right in the middle of the range of predictions. By the way, if you are keeping score, Alan Abramowitz'smodel is presently closest to the two-party breakdown in the national polling. His model generated this prediction based on the president's low approval rating, the second quarter GDP growth, and the fact that Republicans have controlled the White House for 8 years.
Despite the fact that pundits have claimed that Obama is not performing as well as he should be given the economic and political conditions, the models used by political scientists to predict election outcomes--models based on these very conditions--tell a different story. Obama is currently out-pacing the predictions made by some models and lagging only a few percentage points behind others. But his support does not stray more than 4.2% away from any of these predictions. Thus, there isn't much support here for the notion that Obama is greatly underachieving in this election. At least not at this point in the race.
The key takeaway from the recent Pew Report on cell phone only users was not that cell phone only respondents are different, but that even weighting landline only survey data doesn't fully account for excluding cell phone only users. Typically, a survey may be weighted for factors such as age, race, gender, education, and region. This allows pollsters to take a particular sample, and adjust it to look more like what they think the population they are interested in actually looks like. Pew found that even if you weighted a landline sample for all of these factors, that sample still provided results that were 2%-3% less favorable for Obama than one that included cell phone only users.
What exactly makes cell phone only respondents different from those with a landline? If it is simply the case that cell phone only respondents are more likely to be Democrats than those with landlines, then it should be simple enough to correct for not calling cell phones by weighting a sample by party identification. From my reading of the Pew report, they did not examine whether applying a party weight would have accounted for the exclusion of cell phone only respondents. A party weight is something that some pollsters (like Rasmussen) apply, but others do not. However, based on some recent analysis I have conducted using the 2006 Cooperative Congressional Election Study, even weighting by party is not likely to fully account for the differences between cell phone only respondents and those with landlines. There are two reasons for this.
First, it is true that cell phone only respondents are more Democratic than landline respondents. But this relationship is a little more complex than it first seems. In the 2006 survey, cell phone only respondents were just 4% more Democratic than landline respondents and they were 9% less Republican when asked a standard party identification question. However, once you factor in independents who lean towards the Democratic or Republican Party, you find that cell phone only respondents are 10% more Democratic and 12% less Republican. Thus, the party differences are larger when you factor in leaners, a pattern that results because cell phone only respondents are more likely to initially call themselves independents even though they lean Democratic. To fully capture party differences among cell phone only respondents, one would need to factor in leaners.
Second, just looking at party affiliation masks the fact that cell phone only respondents are actually quite a bit more liberal than those with a landline. 35% of cell phone only respondents classified themselves as liberals compared to just 23% of those with a landline. These ideological differences are not completely accounted for by party either. From the table below, you can see that cell phone only Democrats are 10% more liberal than those with landlines. Democratic leaners in the cell phone only sample are 15% more likely to classify themselves as liberal. And even those cell phone only independents who did not express a lean to either party were more likely to be liberal compared to their landline counterparts. Given that cell phone only Democrats and Democratic leaners are more liberal than those with landlines, they should be less likely to defect and vote Republican than landline Democrats.
Thus, this analysis suggests that differences between cell phone only users and those with landlines cannot simply be accounted for by partisanship. In fact, even when I used multivariate models controlling for a wide range of demographic and political factors (party, age, race, gender, income, education, and even religion), cell phone only respondents were still substantially more liberal than those with landlines. Cell phone only respondents are ideologically distinct in ways that cannot be accounted for by party identification or all the other standard demographic factors that pollsters may use to weight samples.
Saturday's Gallup tracking poll revealed two big numbers for Obama. Obama hit 50% in the tracking poll and took an 8% lead over McCain. This isn't the first time that Obama has hit the 50% mark, and it isn't the first time he has held a lead of 8%; but now we are in the last month of the campaign and numbers like these in October usually mean electoral success in November.
Using Gallup's compendium of presidential trial heat polling since 1936, I counted16 candidates who received 50% support or higher in an October Gallup poll. Hitting the 50% mark was a very good predictor of victory. Of those 16 candidates, just two failed to win the general election--Al Gore and Thomas Dewey.
An 8% lead has also been difficult for trailing candidates to overcome. Only one candidate who held a lead of at least 8% in October ended up losing the election--once again, that was Al Gore in 2000. There were thirteen other occasions since 1936 where a candidate had an 8% lead or greater in at least one October poll, and in each case that candidate won.
While the history doesn't look good for Republicans, the McCain campaign can take some solace from 2000. In that campaign, Gore's support reached as high as 51% (and he had a 10% lead over Bush) early in October. However, he quickly lost that lead after the first debate and Bush actually built his own 13% lead in a late October Gallup poll. This was the only time that both candidates reached 50% and held a lead of at least 8% in October Gallup polls--yet another way in which the 2000 election was truly unique. The McCain campaign will have to hope they can duplicate Bush's 2000 comeback rather than go on to defeat like the other 13 candidates who found themselves behind by 8% in October.
NOTE: Mark Blumenthal emailed to point out that the 2000 Gallup tracking poll was highly volatile in 2000 and it may be the case that this volatility was at least partly responsible for both Gore and Bush hitting 50% (and for both taking 8% leads) during October that year.
With Obama taking a clear lead in the national polls with about a month left in the campaign, the question on most minds is whether McCain will be able to make an October comeback and win the election. The problem that McCain faces is that an increasing share of the electorate is committing to one of the candidates at this point, and recent history indicates that few are likely to change their minds.
This is an obvious point, but If McCain is going to get back in this race, he can do so in one of two ways: (1) he can win over undecideds or (2) he can change the minds of those who are currently planning on voting for Obama.
Let's look at the first point. I created the chart below using the super-cool new flash tool that Pollster.com rolled out last week. This chart shows the Pollster.com trend for undecided respondents. At the end of August, the undecided trend was around 10%. By October 1st, that number dropped to just above 5%.
Presently, Obama holds a 5.6% margin over McCain in the Pollster.com trend. Thus, even if you allocated every undecided voter to McCain, it still wouldn't be enough for him to overtake Obama (though it would certainly make for a very close race).
Of course, it is highly unlikely that all of the undecideds will go for McCain, so what about the second option--changing the minds of Obama voters? It turns out that in recent elections, it has been fairly difficult to change peoples' minds in October. The National Election Study conducts a panel survey of voters for each presidential election; they interview respondents face-to-face (no cell phone only problem here) in September and October and then re-interview them after the election. This allows us to get a sense of how common it is for citizens to change their minds in the last month of a campaign. I pulled out the respondents the NES interviewed in September of 2000 and 2004 and the results are in the table below. The columns are the vote preferences expressed by respondents during the September interviews and the rows are the candidates that they actually reported voting for when they were re-interviewed after the election.
In 2004, 94.5% of those who intended to vote for Kerry in September reported having stuck with their choice after the election, compared to 95.7% of those intending to vote for Bush who actually did so. The percentages of those sticking with their candidate are just slightly lower in 2000, but the overwhelming pattern here is that very few voters seem to change their minds in October. In 2004, 4.2% of Kerry supporters changed their minds in October and voted for Bush. Even if McCain manages to get that many defectors, it would only improve his standing by about 2% in the polls (UPDATE: This would also cost Obama 2% in the polls, thereby trimming 4% off Obama's margin). But that assumes that there won't be any defections away from his candidacy. What actually happened in the past two elections is that what few defectors there were largely canceled each other out.
What is also striking is that even people who aren't firmly committed to a candidate appear to end up voting for that candidate in November. The chart below divides respondents into those who said that they were "strong" supporters of their candidate before the election or "not strong" supporters. In the last two elections, at least 80% of the weaker supporters of a candidate stayed with that candidate on election day.
The patterns among individual voters are also evident when you look at the aggregate trends in recent presidential elections. You can see this if you look at the charts from Mark's "convention bump" post in August. In 1980, 2000, and 2004, there was virtually no movement in the polls during the final month of the campaign. In 1988, Bush added a little to his lead in October and in 1996 Dole gained some modest ground on Clinton, but in neither case did the October gains make a difference in the outcome. In 1992, Bush gained significantly on Clinton in October, but attracting supporters from the third party candidacy of Perot may have accounted for some of those gains. In any event, Bush still fell short.
Despite the fact that McCain is only down by 5-7% nationally, time is running out and a comeback seems like a tall order. In the new era of partisan polarization, major October shifts in the presidential polls are unlikely. There are few undecided voters left to persuade at this point and in recent elections we've seen that few voters change their minds once they have settled on a candidate.
There has been a lot of discussion of the Pew Report released earlier this week that shows that including cell-phone only respondents does appear to make a 2-3% difference in the presidential preference polling (see Mark's post). What's most intriguing to me is how this would play out at the state level. Indeed, it seems very unlikely that every state has the same percentage of cell phone only households. Thus, in states with fewer cell-phone only users, the effect of excluding such respondents may have less of an effect on the poll results. On the other hand, states where there are more cell-phone only households may have polling that is further off the mark.
I was hoping to be able to easily find some survey data with enough respondents to get a sense of the prevalence of cell-phone only households in each state. Unfortunately, the 2007 CDC data that is often cited provides more than enough national interviews to accomplish this task, but the dataset hides the state of the respondent, only allowing users to place respondents in a particular Census Region. Nevertheless, we can learn a little about geographical variance from this data. Specifically, families in the South and Midwest are more likely to have cell phones only compared to states in the West and Northeast. Based on this data, we should expect polling from southern and midwestern states to be more prone to error from the cell-phone only problem than polling in other regions.
Another source of data that may be of some use in answering this question is the 2006 Cooperative Congressional Election Study. (In the interests of disclosure, I should say that I am involved in the 2008 version of this study, though I had no role in the 2006 version.) This was a large (approximately 30,000 respondents) internet survey conducted by YouGov/Polimetrix using a matched random sample design. Because this is an internet survey, it probably isn't as ideal for addressing this question as the CDC survey would be. However, the sample was stratified to assure that there would be a large enough sample from each state and since the state of the respondent is available for this data (and isn't for the CDC data), it is the one decent way I've found of breaking out cell-phone only figures by state.
In this survey, 10.6% of respondents indicated that they only had a cell-phone (this is smaller than the percentage cited in the CDC survey, though the CDC survey was conducted a year later). Most interesting is the variation across states. The map below shows this variance.
Some of the swing states that stand out as having higher than average cell-phone only users are Montana (21%), Oregon (17%), Virginia (15.7%), Wisconsin (15.3%) and Minnesota (15.1%). (Keep in mind that these figures are from two years ago). If this survey is providing reasonably accurate figures on cell-phone only users, then it may be the case that polling of these states would be particularly prone to under-stating support for Obama. It may also explain why the polling in some of these states (for example, Virginia) has been so erratic.
Of course, these data may be problematic and should be taken with some caution. If anyone has ideas about other sources that could be used to compile state-by-state measures of cell-phone only households, please let me know. If excluding cell-phone only respondents does matter, then it would be nice to have a strong sense of where it will matter most.
Last week, I outlined the change in party registration figures in a handful of swing states. One of the footnotes to that post was that since Virginia does not have its citizens register by party, such an analysis was more difficult in that state. However, the growth in the registration rolls in Virginia may very well play an important role in this election. Bush won the state by 262,217 votes in 2004; as of September 8th of this year, more than 285,000 people have been added to the registration rolls since that election. The Obama campaign believes that they will see a big payoff from these new voters.
While we don't know the party of the new registrants in Virginia, we do know where those people have been registered and we may be able to draw some conclusions about their likely behavior based on this information. The map below shows the increase (or decrease) in the number of registered voters in each of Virginia's counties since 2004.
The counties with the largest registration increases are located in the northern Virginia suburbs of Washington, DC. Loudon County has added 31,798 new voters to the polls, Fairfax has 25,002 new registrants, and Prince William County has 16,682 more people registered than it did in 2004. (View the top ten counties here: View image). There are also big increases in registration in the counties surrounding Hampton Roads in the southeast of the state and in Chesterfield County, near Richmond.
These figures have made some Democrats very hopeful about Obama's chances in the state. However, it is important to gain some perspective about the extent to which these new registrants will matter in Virginia. For example, look at the types of counties where these new registrants are located.
Over 100,000 new registrations have been logged in counties where Kerry won less than 40% of the two-party vote in 2004 while just 71,466 new registrants are in counties where Kerry beat Bush. Of course, some counties that went for Bush in 2004 may very well go for Obama in 2008 and the Obama campaign may also be registering large numbers of new voters in heavily Republican counties. But it is important to note that the jump in registrants has hardly been concentrated in overwhelmingly Democratic areas. (See a map of the 2004 presidential vote by county here: View image).
We have heard a great deal about the role that Northern Virginia will play in the contest. However, we should recognize that while the increase in registrants in the Northern Virginia suburbs of Washington, DC has been significant--over 100,000 new voters in this area--this is just a drop in the bucket in a state with over 4.8 million registered voters. In 2004, the DC suburbs in Northern Virginia accounted for 32.4% of all registered voters in the state; now, that same area accounts for 32.8% of the state's registered voters. There is no doubt that the growth in Northern Virginia is dramatic, but it takes time to substantially change the demographic and political balance in a state as large as Virginia. Thus, Northern Virginia accounts for just a slightly larger share of the state's electorate than it did four years ago.
Ultimately, it appears that polls are showing a tight campaign in Virginia because a significant share of people who voted for Bush in 2004 are expressing a preference for Obama in this election, not because of the new voters in the state. This is not to say that these new registrants won't matter. While new voters in Virginia won't be the reason that the race close, they could make the difference if the race is close.
One factor that pollsters must deal with is a dynamic electorate. While these changes can be significant at the national level, they may be even more pronounced within particular states. Indeed, we have heard a great deal about the large numbers of citizens that Democrats have registered over the past few years. But how big a difference has this really made? To get a sense of this, I gathered party registration data for seven swing states. The chart below compares the percentage of citizens registering as Democrats and Republicans in 2008 to the same figures in November, 2004 (obtained from the Secretary of State websites).
First, a disclaimer. Party registration is not necessarily comparable across states. For example, New Hampshire's open primary law allows independents to vote in either party's primary while North Carolina's Pennsylvania's closed primary rule means that you must register with a party to vote in that party's primary. Thus, the incentive to register with a party is greater in North Carolina Pennsylvania than it is in New Hampshire (which is why North Carolina's Pennsylvania's figures are higher than in New Hampshire).
Nevertheless, the change (or stability) within each state provides important information and, in most cases, this information appears to favor Democrats. In percentage terms, Republican registration has declined in six of the seven swing states since 2004 (New Hampshire is the exception) while Democratic registration has increased in four of the seven states. Not every state has seen significant partisan changes. The partisan balance in Florida, New Mexico, and North Carolina is quite similar to what it was four years ago. But in four other states, the changes are more significant.
In two states that went for Kerry in 2004 (New Hampshire and Pennsylvania), Democrats have improved their standing among registered voters. In Pennsylvania, a state that the McCain campaign is still targeting, Democrats held a 7% advantage in party registration in 2004 compared to a 12% edge now. Democrats have also made big gains in two states that went for Bush in 2004--Colorado and Nevada. In Colorado, Democrats have cut the Republican registration advantage in half since 2004, from 6% to 3%. But Nevada provides the most stunning example of partisan change. In 2004, Republicans held a 1% edge in party registration; however, just four years later, Democrats now hold a 6% advantage.
These registration figures point to the difficulty that McCain faces in trying to pick up states like Pennsylvania and New Hampshire. Those states went for Kerry four years ago and have only become more Democratic since. The party registration statistics also indicate why Colorado and Nevada have become such good targets for the Obama campaign. In both states, Democrats have made dramatic gains.
What does all this mean for polling? We don't know for sure, but these dynamics do suggest some questions we should be aware of. For example, will these newly registered voters be included in pollsters' likely voter screens? Will those who are newly registered actually turn out on election day? Are all survey firms accounting for these trends when applying partisan (and demographic) weights to their samples? And--the key question--what will the changing partisan landscape in these states mean on election day?
NOTE: See the raw registration numbers here: View image
Unfortunately, states like Michigan and Virginia do not have partisan registration so a similar analysis is not possible in those swing states.
UPDATE: A reader pointed out that in my original post, I mistakenly stated that North Carolina has a closed primary, which is not true. In fact, North Carolina does allow unaffiliated voters to vote in party primaries as long as the parties themselves agree. Both parties have agreed to do this in recent election cycles, so North Carolina has had semi-open primaries in recent years. The states that I've looked at here that have closed primaries include Florida, Nevada, New Mexico, and Pennsylvania.
The latest Survey USA poll out of North Carolina (showing a 20 point McCain advantage) has already generated some disbelief among Pollster.com readers. There may be good reason for some skepticism. After all, less than one month ago, Survey USA showed just a four point lead for McCain. Is this newest poll evidence of a huge Palin bump? An outlier? Evidence of the "shy Tory factor?" Or have Republicans suddenly become much more likely to vote?
I'm not sure we can answer this question definitively yet, but we can lend some context to this poll. At about the same time as Survey USA was releasing the North Carolina result, they were also releasing surveys from Virginia and Washington, neither showing as dramatic a swing. In Washington, Survey USA showed a 7% lead for Obama in early August and now shows him up 4%. In Virginia, McCain held a 1% advantage in early August compared to 2% now. The Virginia result is particularly important since it comes from a state that neighbors North Carolina and shares some demographic similarities. If there was a Palin bounce in North Carolina, presumably we would've seen more of a bump in Virginia as well.
An examination of the cross tabs goes a long way in helping us to understand the North Carolina result. Fortunately, Survey USA provides us with the composition of their sample among various measures. Of particular interest here is their party break down. In early August, 46% of the Survey USA sample was made up of Democrats and just 33% were Republicans. In the most recent poll, the margin was essentially even--40% were Democrats and 41% were Republicans. In other words, in a survey conducted less than a month later, Democrats made up 6% less of the sample and Republican representation increased by 8%.
Everything we know about partisanship suggests that such massive shifts over such a short period are highly unlikely. In other words, it is not very plausible that North Carolinians became 6% less Democratic and 8% more Republican in less than a month. So, what are the other potential explanations? One is the "shy Tory factor" outlined by Nate Silver. According to this theory, it could be that Republicans simply weren't answering surveys at a representative rate a month ago because they lacked enthusiasm, but now they are excited to do so because of the Palin selection. Thus, the increased representation of Republicans in the more recent Survey USA poll may be the result of more Republicans agreeing to be interviewed. But if the "shy Tory factor" is in play, it seems like we'd see a similar increase in Republicans in other surveys. A related theory is that the Palin selection made a lot of Republicans in North Carolina more likely to vote, and since Survey USA presents results from likely voters, they are picking up this change by having more Republicans in their "likely voter" group.
The chart below compares the party composition in the August and September polls conducted by Survey USA in North Carolina, Virginia, and Washington (the numbers don't sum to 100% presumably because one can choose "other" or skip the question).
As indicated in the chart, there were small shifts in the partisan composition of the samples taken in Washington and Virginia, but nothing like what is evident in North Carolina. If there is a "shy Tory" effect or a mobilization effect, it does not appear to be playing out similarly across different states. Most significantly, there appears to be only small (and probably not statistically significant) changes in the composition of the electorate in Virginia, North Carolina's neighbor to the north.
Of course, we really cannot know whether this North Carolina poll is an outlier. But based on the dramatic change in the partisan composition of the North Carolina sample, without a similar change in Virginia, the most likely explanation at this point is that Survey USA may have just drawn a bad sample, one that over-represents Republicans and under-represents Democrats. At the very least, it would be useful if Survey USA can provide more explanation about why they think the partisan composition has changed so significantly in this recent poll.
Just did a little more research and the following two points seem relevant:
According to the excellent North Carolina Board of Elections site, as of September 6th, 45% of those registered to vote in North Carolina are registered as Democrats and 33% are registered as Republicans.
In a July Survey USA poll, the composition of likely voters was 45% Democratic and 37% Republican. (McCain held a 5% lead in that poll).
Give a political scientist a recent economic report and he or she is likely to state with a lot of confidence that Democrats will win this presidential election. Simply put, when the American economy is in bad shape, the out-party tends to have a great deal of success in presidential elections. Yet, less than two months from election day, the out-party's nominee finds himself in a tight race with John McCain. What gives?
Is it simply a matter of the public being distracted? Perhaps the news about Sarah Palin has such weight that it is causing people to forget about the economy? Not likely. Even if the economy isn't affecting everyone personally, the poor unemployment figures released last week coupled with the news of the Fannie and Freddie Mac takeover are likely keeping the economy on peoples' minds. Indeed, in the latest ABC News/Washington Post poll, 37% name the economy as the most important issue, roughly the same percentage as have done so all year. To compare, only 10% named Iraq as the most important issue.
This race isn't close because people have forgotten about the economy, it is close because the public is now largely split on who is better able to handle economic issues. The most recent ABC News/Washington Post poll shows that only a narrow plurality of the public actually thinks that Obama would be better able to handle the economy than McCain (47%-42%). The out-party usually performs well when the economy is doing poorly due to the public's belief that the out-party will do a better job with the economy than the party that was in control while economic conditions worsened. Indeed, in April, the ABC News/Washington Post poll posed the same question, but asked about generic presidential candidates rather than Obama and McCain specifically. The chart below compares these responses.
As the chart indicates, Obama is performing worse on the economy than a generic Democratic candidate while McCain is faring better than the generic Republican. In April, the generic Democratic candidate held a 21% edge on the economy, the kind of dominance on the key issue that most assumed would lead either Obama or Clinton to the White House. Yet, Obama now finds himself holding just a 5% advantage on economic issues and, accordingly, the race between he and McCain is very close.
But Obama's advantage on the economy hasn't always been so narrow, it has been narrowing.The chart below plots the percentage of registered voters saying that Obama or McCain is better able to handle the economy in six ABC News/Washington Post surveys conducted since March.
Note that through July, Obama held a fairly strong (and consistent) lead over McCain on the economy. It is during the past month that his lead has narrowed significantly--from a 17% advantage in July to an 11% lead in August and a 5% edge in the most recent poll. And this narrowing gap is not simply the result of McCain catching up, Obama has lost support on the issue while McCain has been gaining.
What does all this mean for the ultimate outcome? Well, if the economy is as important as political scientists think it is, then vote preferences should track pretty closely with who people think will do a better job on economic issues. The chart below indicates that this is clearly the case.
In March, Obama's lead over McCain on the presidential preference question (13%) was exactly the same as his advantage over McCain on economic issues. Ever since that point, his support has lagged somewhat behind his performance on the economy, but the two have usually moved in the same direction (the one exception is on June 15th, when Obama's economic advantage increased while his overall edge over McCain narrowed). The problem for Obama (and the good news for McCain) is that both have been tracking downward recently.
The McCain campaign has succeeded during the past month in separating evaluations of Bush's handling of the economy from McCain's economic competence. As a result, nearly as many Americans think McCain will do a better job on the economy as think Obama will. It is because they have narrowed the gap on this issue that the McCain campaign is now in a tight race with the Democratic nominee. For its part, the Obama campaign continues to emphasize the economy, recognizing the received wisdom that the out-party should be able to take advantage of poor economic conditions. Yet, this analysis suggests that it is not enough for Democrats to simply direct attention to economic issues--after all, voters are about as likely to pick McCain to handle this issue as they are Obama. If Democrats are going to confirm conventional wisdom, Obama will also need to re-assert the advantage that he held on economic issues when this campaign began.
On Monday, CBS News and the New York Times released a survey of delegates attending the Democratic National Convention in Denver next week. The major finding from this survey came from a question asking delegates who they would like to see as the vice presidential candidate. 28% of the delegates interviewed preferred Hillary Clinton, compared to just 6% who selected Joe Biden (the second most popular choice). The support for Hillary Clinton attracted a lot of attention from news outlets, as well as Pollster.com readers. Of particular interest is whether the 28% figure for Clinton is particularly high or low. To answer this question, we could use a little historical context.
I was able to dig up two earlier Democratic delegate surveys which asked similar questions about VP preferences--one from 1988 and the other from 1992. The results from these surveys are presented in the table below. I've shown the top five finishers in each survey, as well as the percentage naming another candidate ("Other") and the percentage declining to name anyone ("no preference").
The first important point that stands out from this table is that support for Clinton is almost twice as high as it was for any other single candidate in 1988 or 1992. Bill Bradley had the support of 15% of Democratic delegates in 1992, while Jesse Jackson was the preferred candidate for 14% in 1988.
The second notable pattern from 1988 and 1992 is that the eventual VP pick was not one that delegates named in large numbers before the convention. In 1988, only 2% of convention delegates mentioned Lloyd Bentsen as a running mate for Dukakis while just 5% recommended Al Gore in 1992. Thus, Clinton's standing in first place does not necessarily bode well for her chances of ending up on the ticket.
Third, notice that VP ambivalence is not a new phenomenon in 2008, nor is the phenomenon of having a wide swath of politicians named. In 1988, one-third of convention delegates declined to state a preference for Dukakis's running mate, while one-quarter of delegates did not name anyone in 1992. In both years (along with 2008) support was scattered across dozens of names, with only one or two candidates even breaking double-digits in any given year.
Finally, I was able to get the raw data for the 1992 convention delegate survey to look at one additional question: to what extent do a losing candidate's delegates promote their candidate for VP? In 1992, Jerry Brown and Paul Tsongas lost out on the nomination; but each candidate sent plenty of committed delegates to the Democratic convention. The figure below shows who those delegates preferred for Clinton's VP choice.
The figure reveals mixed patterns. Interestingly, Jerry Brown's supporters picked plenty of different possibilities for a Clinton running mate, but almost none of them supported Brown himself. Tsongas's supporters, on the other hand, were more likely to name Tsongas as a potential running mate, but they were also about as likely to name Bill Bradley. For their part, (Bill) Clinton delegates were the most ambivalent, with one-third of them failing to state a preference.
Overall, the historical comparison reveals that Clinton's support is high compared to delegate preferences in 1988 and 1992. However, pre-convention support among delegates didn't do much for Jesse Jackson in 1988 or Bill Bradley in 1992.
As Pollster.com readers have no doubt noticed, there has been much discussion in the posts and the comments here about the merits of polling registered voters (RV) versus likely voters (LV). Mark and Charles have been debating this point in their mostrecentexchanges about whether it is better to include LV or RV results in the Pollster.com poll averages. Charles's last post on this topic raised the following questions:
"There is a valid empirical question still open. Do LV samples more accurately predict election outcomes than do RV samples?"
Ideally, I'd have time to go back over 30 or more years of polling to weigh in on this question. Instead, I thought I'd go back to 2004 and get a sense of how well RV versus LV samples predicted the final outcome. To do this, I used the results from the final national surveys conducted by eight major survey organizations. For each of these eight polls (nearly all of which were conducted during the last three days of October), I tracked down the Bush margin among both RVs and among LVs. The figure below demonstrates the difference in the Bush margin for the LV subset relative to the RV sample from the same survey.
For most polls, LV screens increased Bush's margin, including three surveys (Gallup, Pew, and Newsweek) where Bush did 4 points better among LVs than he did among RVs. But using a LV screen did not always help Bush. In three polls, (CBS/New York Times, Los Angeles Times, and Fox News) his margin remained the same and in the Time poll (which was conducted about a week earlier than the other surveys) Bush actually did 2% worse among LVs.
Of course, this doesn't really tell us which method was more accurate in predicting the general election outcome, just which candidate benefited more from the LV screens. To answer which was more accurate, we can plot each poll's Bush margin among both RVs and LVs to see which came closest to the 2.4% margin that Bush won in the popular vote. This information is presented in the figure below, which includes a dot for each survey along with red lines indicating the actual Bush margin.
Presumably, the best place to be in this plot is where the red lines meet. That would mean that both your RV and LV margins came closest to predicting the eventual outcomes. But, if you are going to be closer to one line over the other, you'd rather be close to the vertical line than the horizontal line. This means that the polling organization's LV screen helped them improve their final prediction over just looking at RVs. If the opposite is true (an organization is closer to the horizontal line than they are to the vertical line), their LV screen actually reduced their predictive accuracy.
The CBS/New York Times poll predicted a 3 point Bush margin for both its RV and LV samples, meaning it was just 6/10ths of a point off regardless of whether they employed their LV screen. Four organizations (Pew, Gallup, and ABC/Washington Post, and Time) increased the accuracy of their predictions by employing the LV screens, coming closer to the vertical line than they do to the horizontal line. Gallup's LV screen appeared to be most successful, since it brought them closest to the actual result (predicting a 2 point victory for Bush despite the fact that their RV sample showed a 2 point advantage for Kerry).
On average, the RV samples for these eight polls predicted a .875 Bush advantage while the LV samples predicted a 2.25 advantage for Bush, remarkably close to the actual result. Of course, this is just one election, but it does appear as though likely voters did a better job of predicting the result in 2004 than registered voters. On the other hand, this analysis reinforces some other concerns about LV screens, the most important of which is the fact that some LV screens created as much as a 4 point difference in an organization's predictions while in three cases LV screens produced no difference at all. It is also important to note that these are LV screens employed at the end of a campaign, not in the middle of the summer, when it is presumably more difficult to distinguish LVs. Ultimately, the debate over LV screens is an important one and the 2008 campaign may very well provide the biggest challenge yet to pollsters trying to model likely voters.
"We have a race that by every measure of every poll is a statistical dead heat. McCain's not supposed to be in this thing, and Obama's supposed to be blowing everybody away and it just isn't happening, at least to this point."
Lou Dobbs (July 17th, Lou Dobbs Tonight)
If you have paid any attention to the news in the past month, you have had a hard time avoiding some journalist or pundit noting that the presidential race is currently a "statistical dead heat" or "essentially tied." The news media, of course, love to cover the horserace aspects of the campaign, particularly in a way that emphasizes how close the election is. But when you step back and gain a little perspective on the big picture, you realize that this race isn't quite the dead heat that it is made to be.
The news media are often a bit myopic in their view of the contest, extrapolating too much from the most recent poll (or even the most recent "surprising" poll). Last week, Fox News released a national survey that showed Obama holding a 41-40% lead, well within the margin of error for the survey. Commentators were quick to emphasize this result and note that the candidates were essentially running neck-and-neck or that the race may even be tied. No doubt there will be a lot of commotion over the latest Gallup/USA Today survey showing McCain ahead (though also within the margin of error) among likely voters. Nevertheless, we gain much better perspective on the state of the race when we look at all available data.
Alan Abramowitz notes that Obama has consistently led in national polls over the past two months. In fact, according to national poll results listed on Pollster.com, Obama had been tied or ahead in 50 consecutive national polls through Sunday. Sure, many polls may show Obama holding a lead within the statistical margin of error, but if Obama and McCain were actually tied, we'd expect as many polls showing McCain ahead as show Obama ahead. Based on some basic calculations, the probability that 50 consecutive national surveys would show Obama tied or ahead if the candidates were actually tied is .0000000000000009. In short, this race is not a "statistical tie," despite what a few scattered surveys (drawing disproportionate attention from the pundits) indicate.