Understanding What Went Wrong with 2016 Polls Will Take Time

Trying to grapple with the failure of polling to predict the Electoral College victory of President-elect Donald Trump, I wrote a short piece for the Castleton student newspaper, The Spartan. In that piece, I argue that we need examine the systematic omission of a segment of the population who are not unreachable, but rather who refuse to participate in polling as respondents. As response rates declined over the past two decades, it was not only a result of those that we could not reach, but we also saw a rise in refusals—those who we could reach but who refused to participate in any polls. Figure 1 shows that this segment of the public is actually greater than the proportion that we cannot reach at all. While we know very little about the unreachable segment of the population, we have some, but limited, information about the refusals. We need to employ that metadata to understand as much as we can about this subpopulation.

The American Association for Public Opinion Research has put together a taskforce to examine the polling from 2016; this is something to watch closely. What I believe that I know now is this:

  1. Any explanation that employs one factor to explain the polling errors is wrong. There are many factors in play.
  2. Most of the early attempts to explain the errors are also wrong; we need a thoughtful and deep examination of the methodology, which will take time and peer discussion.

More to come.

Figure 1. Pew’s response rates, 1997 – 2012.

Thoughts on Sanders and Trump

Sanders and the Democratic party’s nominating process

A NBC News/SurveyMonkey Weekly Election Tracking Poll (May 3- – June 5, 2016) believe that the nominating process is fair; only 38 percent of Democrats label the process as unfair. This doesn’t mean that the Sanders supporters who are upset with the process may come around, but the majority of Democrats are not upset.

What is interesting is that despite the fact that the outsider on the Republican side (Trump) seems to have prevailed, in contrast to the outsider on the Democratic side, the level of Republicans who say that the process for choosing a nominee in their party is unfair is equal to that of the Democrats, at 38 percent. I do not have the raw data, or I would explore who is upset on the GOP side. Are they supporters of an unsuccessful candidate, or are there a number of Donald Trump supporters who still believe the system is unfair despite their having prevailed?

Trump’s attacks journalists and judges

When Trump attacked his political opponents, both within and outside of his party, with demeaning nick names (e.g. “Little Marco” or “Lying Ted”), it seemed sophomoric and unbecoming of any individual seeking the presidency, but it did not raise constitutional questions nor did it seem to threaten the fabric of American democracy, in my mind. However, when Trump questioned the integrity of U.S. District Judge Gonzalo Curiel, the context of Trump’s attacks changed from merely political to potentially threatening to the Constitutional balance among the three branches of government. There is a reason that lawyers can face sanctions for raising unfounded recusal arguments; frivolous recusal requests undermine the system of justice. Imagine if these came from President Donald Trump. To get an idea of the questions that would be raised, simply recall the reaction to President Obama’s derision of the Citizens United ruling during a State of the Union address, and consider that he did not make any charges about the motives or integrity of the justices.

Additionally, many political commentators and political scientists—myself included—have been troubled by Trump’s belligerent tone against the press. Without the benefit of data, I venture to assert that no other presidential candidate has treated the press with as much contempt as has Donald Trump. Gallup’s “Confidence in U.S. Institutions” measure shows that Americans do not have a high level of confidence in the television or print media; only 24 percent have a high level of confidence in newspapers, and 21 percent have a high level of confidence in television news. Interestingly, conservatives have a lower opinion of the media than do liberals, but conservatives have more confidence in television news (21 percent with a high level) than in newspapers (16 percent). (See Gallup. “Trust Differs Most by Ideology for Church, Police, and Presidency.”)

The public’s low confidence in the media makes Trump’s hostility to the press easier to understand. And yet the press has repaid Trump’s hostility with a wealth of earned (free or nonpaid) media coverage. According to the New York Times (March 15, 2016), Trump had received over $1.8 billion in earned media by that point in the campaign, eclipsing all other candidates by far. In fact, the second highest level of earned media went to Hillary Clinton, $746 million, or less than 40 percent of what Trump received.

The days of chumming with the press like John F. Kennedy did are long gone, and this has been to the electorate’s advantage. It would not surprise me to learn that most candidates carry disdain for the press, but candidates, as a rule, do not behave in an openly hostile manner to the press. Most candidates understand that the press—the fourth estate—is necessary to keep the public informed, to be a watchdog for public interests, and to provide voters with the information they need to make intelligent decisions in the voting booth.

Sources of Trump’s Support

At the 2016 New England Political Science annual conference in Newport, RI, in April, Matthew MacWilliams from UMass Amherst presented his research in a presentation titled, “The Rise of Donald Trump: America’s Authoritarian Spring.” The provocative title drew political scientists into the panel like moths to a light on a summer’s night, and the presentation did not disappoint. MacWilliams, who is the President of his own political consulting firm called MacWilliams Sanders Communication, used survey data to measure respondents’ support for authoritarianism in a multivariate scale. Among the measures used in the scale are support for suspending habeas corpus, support for “keeping other groups in their place,” and preventing minority opposition. Those who scored highest on the authoritarianism scale were more likely to support Trump than those who scored low.

In fact, when controlling for several variables, such as age, education, ideology, and religion, one’s authoritarian score was the best predictor of one’s support for Trump. In an article for Politico (January 17, 2016), MacWilliams wrote,

So, those who say a Trump presidency “can’t happen here” should check their conventional wisdom at the door. The candidate has confounded conventional expectations this primary season because those expectations are based on an oversimplified caricature of the electorate in general and his supporters in particular. Conditions are ripe for an authoritarian leader to emerge. Trump is seizing the opportunity. And the institutions—from the Republican Party to the press—that are supposed to guard against what James Madison called “the infection of violent passions” among the people have either been cowed by Trump’s bluster or are asleep on the job.

We have been warned.

Here’s Looking at You, South Carolina

Billionaire and Republican candidate Donald Trump took all of the 44 delegates at stake in the February 20, 2016 Republican presidential primary in South Carolina. Trump’s victory was projected early in the evening, yet the media continued to closely follow the returns to ask, “who took second place?” They also may have asked, “who did not get any delegates,” or “who, between two candidates in a near tie, barely edged out the other to gain no delegates?” (See previous blog posts.)

So I asked in a tweet, “What does it matter at this point who takes second place?” One response I received was that it mattered to Bill Clinton, who rode a second place finish in New Hampshire in 1992 to become the nominee. Clinton in 1992 is, in fact, a perfect case in point. The media set the expectations for candidates, and then when candidates surpass those expectations, the media praise the candidates for doing better than we thought they should. Rather than accepting that the expectations were misguided, the media suggest that the candidates have excelled. So a second place finish becomes a victory.

And in the 2016 South Carolina Republican Primary, three candidates gave victory speeches despite the fact that only one of them won any of the state’s 44 pledged delegates. Trump had a plurality in each of the state’s congressional districts as well as a plurality state-wide, and the winner-take-all at the district level meant that there were no delegates to be had for a second-place finish. Yet Senators Marco Rubio and Ted Cruz both made speeches that sounded triumphant, and the media remained intently interested in which one of these two would edge out the other when all votes were counted.


Because the narrative has to be written in broad strokes and cannot accept a little bit of nuance that would suggest such a narrow margin is indicative of no larger trend. Just look at the New York Times report on the outcome; without a narrative, it’s hard to care who is in second place.

Perhaps if we saw these contests more from the voters’ perspective than from the candidates’ points of view, we could recognize that while a plurality of South Carolina voters, evenly distributed throughout the state, favor the outsider candidacy of Donald Trump, the remaining voters—making up a majority in total—are not settled on any candidate, and both Rubio and Cruz are about equally attractive to the GOP primary voters.

Adieu to Jeb!

Having spent so much air time, twitter time, and ink on the 0.2 percentage points separating Rubio and Cruz, the media hardly mentioned that former Florida Governor Jeb Bush, who suspended his campaign last night, edged out Kasich and Carson by margins not much different than that between the two senators.

Everything has Changed Upon the Death of Antonin Scalia

The United States Supreme Court building

The death of Supreme Court Justice Antonin Scalia will inevitably have a major impact on the 2016 presidential race. Given that so many Court decisions have been made by a 5-4 division of the Justices and that Justice Scalia is the conservative intellectual leader of the bench, the balance of the Court is now in play.

Scenario One:

President Obama nominates a candidate and the Republican-led Senate confirms the nomination. For this unlikely scenario to play out, the nomination would be a consensus candidate, which—like the Yeti and the Lock Ness Monster—are fun to imagine but may not exist in reality.

Scenario Two:

President Obama nominates a candidate and the Senate does not move to confirm. This scenario is very likely, kicking the can down the street to the next administration. The Republicans in the Senate have no incentive to give the Court over to the other side, and are likely to wait and see if the Republican presidential nominees wins in November. Even if the GOP nominee loses the race for the White House, the Republicans have lost nothing in terms of the Supreme Court that they wouldn’t have lost otherwise. The Senate can then consider if Obama’s nominee is likely to be better or worse than the nominee they are likely to get when the next president takes office.

Either way, future debates, stump speeches, and votes will likely focus on the new political landscape. After all, presidents only hold office for 4 years—eight years if they are fortunate—but a federal justice could shape the U.S. government for three to four decades.

Fewer Candidates, but the Same Story

Here is a brief post wherein I double down on my conviction that the presidential nomination races, in both parties, are simply a race to amass a majority of delegates, not a series of beauty contests with the objective of revealing to the party faithful which candidate is the most electable in the general election. Let’s be clear that when party leaders chose the nominees and very few delegates were determined by the primaries, the old narrative fit. But the nature of the caucuses and primaries have changed, and the conventions have become the time where the awards of the state-level contests are handed out—not a time to consider what happened in the states and then make a choice. The choices have been made by the time of the conventions.

So in the Republican nomination battle, the number of candidates have winnowed from an unwieldy 17 down to a manageable six (with the recent departure of Jim Gilmore, former Governor of swing state Virginia, who not only was never invited to participate in a major debate but also was excluded from the so-called undercard debates). Despite the reduced number of players, the same fissures remain: there are the outsiders (Trump and Carson), the candidate who where’s the “conservative” label while attacking the party structure (Cruz), and the candidates who are fighting for the party establishment (Rubio, Bush, and Kasich).

On the Democratic side, it has always been a the inevitable (Clinton) against the outsider (Sanders)—reform versus revolution. In the minds of many on the Democratic party’s periphery, it is the competent but untrustworthy candidate with the right resume versus the authentic, idealistic champion of those who are not among the party elite. So Sanders wins support among Vermont voters, while all of Vermont’s highest-level elected officials throw their support behind Clinton.

For both parties, these are the conflicts that were extant at the start of the nomination race, and they remain today. This does not change until one candidate can eliminate the representative of one of the major fissures of conflict. Until then, it’s all about collecting delegates.

Project 240 Mock Primary Results

On February 9, 2016, Project 240 (a collaboration between the Paramount Theatre and Castleton University), hosted a Mock Primary designed to coincide with the New Hampshire primary.  Those in attendance were treated to a showing of the 1960 documentary Primary, including commentary and discussion by Castleton faculty Michael Talbott and Rich Clark, a discussion by Senator Pollina about the date of Vermont’s primary, a presentation of student-collected data about the political process gathered by Assistant Professor Jennifer Turchi’s Sociology Research Methods class, and an opportunity for everyone in attendance to cast a ballot in the Project 240 Mock Primary.

Votes were tallied and presented during the event. Here is a summary of the Project 240 Mock Primary results:

  • A total of 89 votes were cast with 63 attendees selecting the Democratic ballot and 26 selecting the Republican ballot.
  • Bernie Sanders won the Democratic vote.
Project 240 Mock Primary Democratic Results
Project 240 Mock Primary Democratic Results
  • In a close race on the Republican side, John Kasich won the mock primary by one vote over Donald Trump.
Project 240 Mock Primary Republican Results
Project 240 Mock Primary Republican Results

The next Project 240 event will be on February 28th with a fantastic evening of music.  Two more primary debates are on the Project 240 schedule on March 9th and 10th. (For more information go to: http://project240.org/)  We hope you’ll join us for these great community events as we make our way through the election season!

It’s All About the Delegates

Given the focus on opinion polls—a topic which I am thoroughly immersed—it is nevertheless important that we remember that the major party presidential nominees are neither chosen by polling results, nor by plurality elections in the 50 states. Rather, the nominees will be chosen in July by party delegates gathered in the respective conventions—in Cleveland for the Republicans and in Philadelphia for the Democrats. The primaries, caucuses, and in a few cases conventions, will determine which delegates go to the conventions.

Now to correct myself, the 39 primaries, 14 caucuses, and 3 conventions will select 93 percent of the Republican delegates and 85 percent of the Democratic candidates. In each state and territory, the top three Republicans will automatically go to Cleveland regardless of the state votes. On the Democratic side, 713 party leaders and elected officials (PLEOs), or 15 percent of the 4,764 Democratic Party delegates, will go to Philadelphia regardless of anyone’s vote; these are called the Super Delegates, and they are unbound by any popular will and free to vote their conscience.

GOP Delegate allocation
How GOP Delegates are allocated by the national party

The rules for allocating the number of delegates to the various states and territories are made by the national parties. The rules for selecting those delegates are made by the state parties. There is no single entity determining how the process works.

In both parties, the nominee needs to obtain a majority of the delegate votes at the convention. If no candidate, after the first round of voting, receives a majority of the vote, the delegates continue to vote until a point where someone—and it need not be from among the candidates in the primaries or caucuses—receives a majority of delegate votes. (It is the dream of every political scientist to witness a brokered convention, where the selection of the nominee plays out on live television through bargaining on the floor of the convention.)

Before March 1st, which is being called Super Tuesday or SEC Tuesday to reflect the large number of delegates at stake and the large number coming from southern state, only 133 of the 2,472 GOP delegates (5 percent) are in play. I do not mean to deny the importance of these primaries and caucuses, but if a candidate were to sweep these contests, they would still be a very long way off from securing the nomination.

As a rule, early states allocate delegates by various versions of proportional representation, although South Carolina a winner-take-all statewide and at the district level. After the big Super Tuesday contests, only 822 (or one third of the votes) will be selected in the Republican nomination process. Consequently, even is one GOP candidate rises far above the pack, they will still have to amass another 415 delegates in order to have the barest of majorities and a thin claim on the nomination (thin only in that they would need to hang on to each and every delegate).

By March 15, a majority of GOP delegates will have been selected, although the large GOP field will make it very unlikely that any candidate would have secured a majority by this point. It is likely that no Republican candidate will have the majority until at least April, and it is very possible that we won’t know the Republican nominee for sure until the June 7th, winner-take-all California primary. It is also possible that no GOP candidate has a majority of the pledged delegates going into the Cleveland Convention July 18, 2016—a political scientist can dream.

GOP Delegates large and small
Number of GOP delegates by state, largest and smallest

I will say more on the Democratic race in another blog piece.

What the National Republican Party Can Learn from Vermont: a response to David Brooks

In his November 13, 2015, New York Times column, David Brooks wrote about the future of the GOP and the party’s positions on immigration. Brooks writes, “The demographics of this country are changing. This will be the last presidential election cycle in which the G.O.P., in its current form, has even a shot at winning the White House.” The GOP’s base is older and far less racially and ethnically diverse than the population as a whole. So Brooks begins his column with the question for his fellow Republicans, “Are we as a party willing to champion the new America that is inexorably rising around us, or are we the receding roar of an old America that is never coming back?” (New York Times online, November 13, 2015).

Like the Republican base, Vermont as a state is older and aging and is far less diverse than the rest of America. And yet, Vermonters—Democrats, Republicans, and Independents alike—would likely respond to Brooks’ question that they will champion the new America. As Vermont Governor Peter Shumlin reaffirms the state’s willingness to accept refugees from Syria, even while some of his fellow governors are expressing obstinate refusals, Vermonters generally express openness to immigrants, minorities, and, to use the colloquial term, flat-landers from everywhere.

Vermont was an early adopter of same-sex unions. In a June 2013 poll by the Castleton Polling Institute, 66 percent of Vermonters expressed the view that same-sex couples should have the right to marry, and 74 percent thought that same-sex couples should receive the same federal benefits as heterosexual couples. These views represent a sense of comfort with, rather than reaction against, changing times.

A Seven Days article from January 15, 2014, discussed the process of resettling refugees from all over the world in Chittenden County (Kevin J. Kelley, “Twenty-Five Years and 6,300 People Later: A Vermont Refugee Report”). The article proclaims that the settlement of people from Africa and Asia into the largely homogeneous communities of Vermont had come about with very little xenophobic or reactionary activity. “Vermont is regarded as such a welcoming place that many immigrants move here from elsewhere in the U.S.,” notes the deputy director of the Association of Africans Living in Vermont, according to the article.

In a recent talk at Castleton University (October 13, 2015), Lieutenant Governor and gubernatorial candidate Phil Scott suggested that Vermont would benefit from an influx of immigrants who are younger than our average age. On that same page, Governor Jim Douglas has said that at the base of the problems facing Vermont is demographics. The same can be said of the GOP, but the responses to what Brooks calls “the new America” from GOP presidential contenders, contrasts significantly with the responses from Vermonters who embrace the new.

Demographics are not determinant.

Vermont’s Party Politics: a model for the nation

To the rest of the country, Vermont is a liberal, hippie—bordering on socialist—enclave of New England. If Americans know anything about Vermont apart from Ben and Jerry’s and Cabot cheeses, it is that we have pushed the envelope for liberal causes: same-sex marriage, single payer healthcare, and now a flirtation with legalizing recreational marijuana, not by a referendum but by the standard legislative process.

The politicos of the nation will know Vermont for its current congressional delegation, including an establishment Democrat and an Independent in the Senate and an affable representative willing to work across party lines in the House. In other words, those in the know have a more balanced sense of Vermont politics. We have sent to Washington moderate Republicans such as Robert Stafford and James Jeffords. And when the political winds changed, these legislators stayed moored to their moderate positions.

While to the rest of the nation, Vermont appears to be a solidly blue state, we have since 1962 had competitive two-party battles for the highest executive office, with the current governor winning twice (of three elections) by a very slim margin, and neither time carrying a majority of the popular vote.

Vermont is a two-party state. Republicans may be in the minority, but they are an active minority with a potential of taking back the reins of government.

In a recent poll by Castleton University’s polling institute, the highest-ranking Republican in the state, Lieutenant Governor Phil Scott, received the highest approval rating overall with 71 percent rating him favorably and only 23 percent unfavorably. Among those of his own party, Scott’s numbers are fabulous: 77 percent favorable and only 4 percent unfavorable. Yet even among Democrats and Independents who lean Democrat, Scott’s favorables are 66 percent to just 12 percent unfavorable (an astonishing 54 percent net favorability rating).

In addition, Scott has pledged to run a campaign focused on the state’s economy, which a plurality of Vermonters cite as the most important problem facing the state.

Scott’s popularity, as well as the continued strength of GOP gubernatorial candidates, is no fluke. In his memoir, The Vermont Way, former-Governor Jim Douglas argues for civility in politics and focusing on issues over partisan advantage. Commenting on the hyper-partisanship in Washington, Douglas writes, “It is profoundly unfortunate, not only because most Americans, I believe, don’t care much for the political extremes and would rather see elected officials achieve results, but also because blind allegiance to party does not lead to the policies that make for a stronger, more prosperous nation” (p. 324).

Phil Scott is clearly within the Douglas mold; he has a reputation for putting policy above politics, and he has earned respect from both sides of the aisle for this practice, as our poll results indicate. As the leader of the minority party, this bodes well for those who wish to move beyond partisanship.

Nationally, Democrats and Republicans have clear differences in what they believe the role of government should be. And clearly, these differences run deep. It is good for America to have competitive parties—to have clear choices in the direction of the nation—but because we do not have a parliamentary system where one party gets its chance to control the entire government for a period of time, once elections have played out, the parties must work together enough to govern effectively. This is what we do in Vermont largely because the Vermont Republicans continue to govern even in the minority. They not given up on government, as so many have at the national level, and they have continued to participate and make practical changes where possible.

Although Vermont is not without partisan bickering, and as Douglas notes in his memoir, partisanship is on the rise in the Green Mountain state, it is nowhere near the level seen in the nation’s capital. In large part, that is due in part to GOP leaders in Vermont who have not turned their backs on government for partisan advantage or ideological purity and to Vermont voters who continue to look past partisanship in selecting leaders.

The Vermont Way–or the Tao of Vermont–is a lesson for the nation.

Musings about the Republican presidential nomination race

Despite the large, crowded field of candidates for the 2016 Republican presidential nomination, it is very possible that the ultimate choice has not yet officially entered the race. While the story for the Democratic nomination race is about challenges to the inevitable nominee, Hillary Clinton, the story on the Republican side is a wide-open race, with no inevitability despite some of the high profile Governors and Senators in the race.

The great challenge for the Republican candidates is to appeal to Republican primary voters without alienating or upsetting independent voters in swing states. In other words, it is more difficult for a candidate with deep appeal to Republican Party conservatives to position themselves on major policy issues where they are likely to appeal to a majority of general election voters. Recognizing this dilemma, Jeb Bush has proposed running a primary campaign that is no different from a general election strategy; in this vein, his policy positions have tended to be more moderate in the area of immigration, and he has demonstrated on many occasion his bi-lingual skills. On the other hand, Scott Walker has vowed to hold true to conservative principals and not moderate his views for a national audience.

At this time, the Real Clear Politics polling averages, at the national level, show Bush, with only 15 percent, holding a slight lead over the rest of the pack, but Donald Trump has seen a surge in recent polls by CNN and Fox News. Every candidate in this crowded field needs to separate themselves from the pack somehow; for Trump, his resume alone separates him from a long list of governors (current and former) and senators, but he has also made some inflammatory statements that have served to keep him the focus of major media outlets, for good or ill. For the other candidates, the debates may provide them the opportunity to set themselves apart from the field, provided that they make the cut to appear in debates.

While I am an avid followers of the polls, I’m not sure that the polling data have much to offer at this time in terms of predicting who the Republican nominee will be. If one factors margins of errors in the estimates, there may be several candidates with a lead among likely Republican primary voters. So we need to consider what other factors may allow a candidate to continue their bids for the nomination after Iowa and New Hampshire (and then again after South Carolina and Nevada). The silent primary—funds raised from contributors—is one data point that will separate the candidates. Although Jeb Bush leads the pack in the total raised to this point, Scott Walker has not been a candidate, officially, and has not had to declare. Of those who were declared candidates, Ted Cruz and Marco Rubio are just behind Bush. Apart from fund-raising, we can assess the networks built and the national reputation of the candidates. Based on this, I predict that Bush, Cruz, Rubio, Rand Paul, John Kasich, and Walker will still be viable candidates after the Nevada caucus. It’s likely that a few of the middle-tier candidates will also still be running, such as Ben Carson, Rick Santorum, and Rick Perry. Consequently, the Iowa caucus and New Hampshire primary will not likely have the winnowing effect that they’ve had in the past. It’s going to be a long ride.


A quick methodological note:

The horse race question in the CNN/ORC poll read, “I’m going to read a list of people who may be running in the Republican primaries for president in 2016. After I read all the names, please tell me which of those candidates you would be most likely to support for the Republican nomination for president in 2016, or if you would support someone else. Jeb Bush, Ben Carson, Chris Christie, Ted Cruz, Carly Fiorina, Lindsey Graham, Mike Huckabee, Bobby Jindal, John Kasich, George Pataki, Rand Paul, Rick Perry, Marco Rubio, Rick Santorum, Donald Trump or Scott Walker.” Of course, keeping with best practices, the order of the candidates’ names were randomized, but even with this, there has to be some concern for the recency effect—the propensity to choose the last option from a long list because it is the easiest to recall.