The “Generic” Congressional Ballot Question

Democrats would lead by nearly fourteen points in “generic” Congressional ballot polls next November if the trends seen since Trump took office continue.

I have written earlier about how methodological differences among pollsters can lead to significantly different results.  In my analyses of Presidential approval I showed how Donald Trump’s approval ratings varied depending on the choice of sample to interview and the interviewing method chosen.  In this piece I apply the same approach to the so-called “generic” ballot question, typically “If the elections for Congress were being held today, which party’s candidate would you vote for in your Congressional district?”  Some pollsters mention the Democrats and Republicans in this question, others leave it more open-ended like the example I just gave.

I have focused on the net difference in support for generic Democratic and Republican candidates.  This ranges from a value of -4 (Republican support being four points greater than Democratic support) to a high of +18 in the Democrats’ direction.  Here is a simple time plot showing how support for the Democrats on this question has grown while Trump has held office.

The Democrats held a small lead of just under four points on the day Trump took office.  Since then the Democrats’ lead has slowly increased to an average of eight points.

What’s surprising about these data is that they do not show the usual methodological differences we see in the presidential series.  Here are a few regression experiments using my standard array of predictors.

Choice of polling method has no systematic relationship with Democratic support on the generic ballot question. In contrast, Trump’s job-approval ratings run one to two points higher in polls taken over the Internet.  Another striking difference is the greater level of support found for Democrats in polls of registered or “likely” voters.  Again, the job-approval polls show an opposite effect, with polls of voters displaying greater levels of support for Trump than polls that include all adults.  I have also included separate effect measures for the two most-common pollsters in this sample, Politico/Morning Consult and YouGov/Economist.  Job-approval polls taken by the former organization show a pro-Trump “bias” of about three percent; on the generic ballot their polls place Republican support about five points higher than other polls.  YouGov/Economist polls also have Republican tilt on this question, though they show a slight anti-Trump bias in job-approval polls.

If we extrapolate these results to the fall election on November 6th (655 days after the Inauguration), and include the effect for registered voters, the model predicts the Democrats’ lead in generic ballot polls would reach nearly fourteen percent (=4.07+2.62+0.011*655).  A margin that large would easily overwhelm the built-in advantage Republicans hold based on partisan self-selection and gerrymandering.  Even if the Politico figure is correct, adding in that pro-Republican factor brings Democratic support down to nine points on election day.  That result would still reach nine percent, or a Democratic/Republican split of about 54-45.  That 54 percent figure still exceeds the 53 percent minimum I estimated earlier would result in Democratic control of the House of Representatives.

Using the model for the relationship between seat and vote divisions presented earlier, a 57 percent margin in the national Congressional vote would translate into the Democrats’ winning 55 percent of the House seats for a margin of 239-196.

Republicans Continue to Leave the House

Three more Republican House Members announced they would not seek re-election this week, bringing the total number of retiring Republican Members to 34 according to the New York Times. That figure compares to 16 Democrats, for a net Republican difference of +18.  We have to look back to the Democratic landslide in 1958 to see a mid-term with double-digit net Republican retirements.  For Democrats, only in 1938 and 1978 did the number of their retirements exceed Republican retirements by ten or more.

This increase of three net Republican retirements raises the predicted Democratic seat swing to 41 using the relationship depicted in the previous article.

I have shown in earlier postings that the relationship between seats and votes that advantaged Democrats in the years after World War II moved steadily in the Republicans’ direction beginning in 1980 and, with the help of gerrymandering, became even more favorable for the Republicans after 2010.  That might temper our belief in a prediction for an election being held in 2018.  First, the 2018 retirement margin of -18 is close to the observed maximum of -21 in 1958.  Perhaps the 1958 election is an “outlier” and without it the relationship is less steep than we observe.  However slope and intercept coefficients estimated with 1958 excluded are numerically nearly identical to those estimated with that year included.  So it’s unlikely that the historical model is radically overestimating the likely result next fall.

Another test is to let the relationship differ before and after 1992 to see whether the structural changes that we observe in the seats/votes relationship in the current era appear in the relationship for retirements and seat swings.  Once again, allowing the coefficients to differ before and after 1992 showed no measurable statistical difference. While the effects of gerrymandering and partisan self-segregation may make the House less vulnerable to “waves” of Democratic support, there is no evidence for that thesis looking at retirements as a predictor of seat outcomes.

These estimates have a lot of uncertainty attached.  The standard error of estimate is about 28 seats.  That means there is about a two-thirds chance that the actual swing will be somewhere between 13 and 67 seats.  Since the Democrats need a swing of at least 24 seats to win control of the House, even a retirement margin of 18 is not enough to ensure a change in party control.

The regression model taking President’s partisanship into account is a bit more conservative; it predicts a swing of 37 seats.

Update (2/26/18) – One more Republican has announced he is leaving the House, along with one more Democrat.  The net difference remains at +18.

What Do House Retirements Tell Us About the Future?

The pace of Republican retirements predicts that Democrats should take back control of the House of Representatives this fall by a margin of eight or nine seats.

This week Edward Royce (R-CA), Chair of the House Foreign Affairs Committee, and Darrell Issa (R-CA), former Chair of House Oversight and Reform, joined 27 of their fellow House Republicans by announcing that they are retiring from the chamber.  About half are leaving public office entirely, while the remainder are seeking another office like governor or senator.  (Issa has threatened to run for the House again in an adjacent district.)

On the other side of the aisle, fourteen Democrats have announced that they will be leaving the House of Representatives.  Only five of them are leaving public office, though Ruben Kihuen (D-NV) may be joining them subject to an investigation into allegations of sexual harassment.

Many pundits have interpreted the much higher retirement rate for Republicans to be a bellwether for this fall’s Congressional election.  If no one else announces a retirement, the Republicans will face a net loss of fifteen seats going into the midterms.  Just how large a threat do such retirements pose to Republican control of the House?

These two charts show the relationship between the net Republican margin of victory in terms of seats in the House and the net partisan difference in retirements.  It turns out that retirements tell us essentially nothing about Congressional outcomes in Presidential years, but they are quite informative in mid-term elections like 2018.  Here is the chart for Presidential years:

In presidential years we see little relationship between net Republican retirements and how well the party fares in the upcoming general election. What matters more are Presidential “coattails” with Republican swings in years when Eisenhower (1952), Nixon (1960, 1972), and Reagan (1980) ran.  Democrats were favored when they ran along side Johnson in 1964, Obama in 2008, and Franklin Roosevelt in 1944.

A much different picture appears if we look at the same relationship for mid-term years.

Now the number of Republican seats won or lost depends much more directly on the number of retirements.  The line is anchored by the Republicans’ success in the 1938 midterm and their dramatic losses in the 1958 election. If the figure for net retirements remains at fifteen, the Republicans are predicted to lose about thirty-two seats next November.  That would give the Democrats control of the House with a margin of eight seats.

However, because the President’s party historically loses seats in midterms, we should expect to see more retirements from the President’s party in midterm years.  When Democrats presided over a midterm, an average of four more Democrats retired from office than did Republicans.  When the Republicans held the White House in a midterm year, retirements from their ranks outnumbered Democratic retirements by an average of six.

So some of the strong relationship we see between retirements and midterm losses arises simply because the President’s co-partisans are jumping ship knowing that their party will do more poorly in the upcoming midterm.  That leaves us with the question of whether retirements have any additional predictive power once we take the President’s partisanship into account.  Retirements still matter in this better specification, though that effect just achieves statistical significance.

Based on these estimates, in a year where the Democrats hold the White House, and the number of retirements on both sides of the aisle is equal, the model predicts that the Republicans should gain about thirty seats.  When the Republicans hold the White House, and retirements are equal, the Democrats should gain about twelve seats (= 30.2 – 42.1 = 11.9).  Regardless of which party controls the Presidency,* Republicans are predicted to lose 1.4 seats for every retirement.  Applied to the current circumstances, this formula predicts a Democratic victory by thirty-three seats, or one more than predicted by the simpler model.



*Allowing the relationship between retirements and seat outcomes to vary separately depending on which party controlled the White House added no explanatory power.


How Doug Jones Won

Comparing last night’s results for the Special Election in Alabama to prior elections in that state shows the path which brought Doug Jones his unlikely victory.  Like all special elections, the Alabama election unsurprisingly failed to mobilize as many voters as last year’s Presidential contest, but the turnout last night did exceed the 2014 mid-term figure by four percentage points.  Alabamians apparently thought this race was worth the effort.

Jones gained nearly as many votes in this special election as Hillary Clinton polled in November, 2016, despite an electorate 37 percent smaller than for the Presidential election.

Doug Jones expanded the Democratic electorate in the largest Alabama counties when we compare his results to those polled by 2014 Democratic gubernatorial candidate Parker Griffith.  (There was no Democratic candidate for Senator on the ballot that year.)

Jones won 671,151 votes last night, an improvement of 57 percent statewide over Griffith’s 427,787 total three years ago.  The graph shows clearly that Jones’s advantage grew as a function of county size as measured by total turnout.  His best performances were in Madison County, where aerospace center Huntsville is located; Shelby County, which includes suburban Birmingham; and Baldwin County, just east of Mobile.  Moore carried both of the latter counties over Jones but by severely diminished margins compared to Republican performance in prior elections. Jones doubled the Democratic vote in Lee County, home to Auburn University, and made a significant gain in Tuscaloosa where the University of Alabama is located.

Moore’s support, in contrast, was strongest in the smallest counties.  Here I am comparing Moore’s performance last night to the total vote cast in the Republican primary runoff election against Luther Strange late last September.  Moore needed to mobilize sufficient numbers of Strange voters to add to his own totals going into last night’s election.  He failed to do so.

Statewide Moore received about 35 percent more votes than he and Strange together polled in September.  However Moore saw his smallest gains in the largest counties, the opposite of the pattern we saw for Jones.

Running a simple regression of Jones’s lead over Moore against demographic variables shows the dominant power of mobilized African-Americans with smaller effects for the proportions of Hispanics and people with a college degree.

In a county with no blacks, no Hispanics, and no one with a college degree, Moore would beat Jones by 79 points, e.g., 89-10.  The size of the black population played the most important role in determining support for Jones.  His lead expanded by 1.6 percent for every one percent increase in the percentage of blacks.  Hispanics played a less significant role with a coefficient about half the size of that for blacks that barely achieves statistical significance.  The proportion of a county’s residents holding a college degree mattered nearly as much as the size of its black population. These three variables alone account for about ninety-four percent of the variance in Jones’s lead over Moore.

Jones’s victory with a margin of 1.5 percent over Moore suggests that the polls that included cell phone respondents were right on track.  As I wrote yesterday, “Taken together, [my analysis of polling data] suggests that Jones has averaged a 1-2 percent lead in polls taken since the Washington Post story that included calls to cell-phone users.”  Polls limited to landline households, which predicted a Moore victory, were off the mark.


Some Observations on the Polling in Alabama

Alabama’s citizens will head to the polls to vote in the most competitive Senate election that state has seen for decades.  Democrat Doug Jones is trying to wrest the state into his party’s column while Republicans rally behind Roy Moore.  Most anyone reading this blog knows about the issues in this race, so I’m going to focus solely on the polls as archived at RealClearPolitics.  Since RCP does not publish data on polling methods, I examined the individual reports for each poll and, when the method used was unclear, contacted the polling agencies directly.

Polling results in this race have shown little convergence as we approach election day.  The Republican dominance in Alabama’s elections has meant that few national polling agencies have paid much attention to the state over the recent elections.  As a result, few national polling organizations have much experience surveying Alabama’s voters.  That has changed a bit as the election achieved national prominence, but still the vast majority of Alabama polls come from organizations will limited track records.  Over at FiveThirtyEight, Harry Enten observes that “Alabama polls have been volatile, this is an off-cycle special election with difficult-to-predict turnout, and there haven’t been many top-quality pollsters surveying the Alabama race.”

“Top-quality” pollsters rely on live interviewers making calls to both landline phones and cell phones.  FiveThirtyEight adds the additional criterion that the polling agency be involved with national organizations like the American Association for Public Opinion Research.  I will limit my analysis to just whether calls were made to a sample of cell phone owners.  As it turns out, this factor alone has a profound effect on a poll’s estimated margin between the candidates.

Here is a list of the available polls based on their method of interview.

Only one poll among those that included interviews with respondents via cell phone shows a lead for Moore; in contrast, only one of the landline-only polls puts Jones ahead.  The “swing” is quite substantial, about an eight-point differential based on the method used.

This difference arises from the much higher usage of cell phones by younger respondents who prefer Jones in most polls.  For instance, in today’s poll from Fox News likely voters under 45 year of age preferred Jones 59% to 28%, while voters above that age preferred Jones by only a one point margin, 45% to 44%.

I also modeled the difference in support between Jones and Moore using my standard predictors, time left before the election, and dummy variables for polling methods.  I also added a dummy which is coded one beginning on November 9th when the story about Moore’s alleged molestation of young girls was released in the Washington Post.  The variable measuring proximity to the election proved statistically insignificant, leaving just three dummies, whether the pollster made calls to cell phones, whether live interviewers were used, and whether the story had broken in the Post.

In polls taken before the publication of the molestation story, Jones trailed Moore by an average of eleven points.  Since then Jones has seen an average gain of six points, not enough on its own to return the race to even.  However polls that interviewed respondents via cell phones show a slightly larger difference of nearly seven points in Jones’s favor.  Taken together, these results suggest that Jones has averaged a 1-2 percent lead in polls taken since the Washington Post story that included calls to cell-phone users. (Update: Jones’s margin of victory over Moore was 1.5 percent statewide, right in line with this prediction.)

I also included a term for whether live interviewers were used.  Since all polls that include cell phone owners must use live interviewers by law, this remaining group represents organizations that polled only landline owners with human interviewers.  I find a small, though statistically insignificant (p<0.17) positive effect on Jones’s support from people surveyed by live interviewers.  It is hard to interpret what this effect might signify.  It could represent an unwillingness among Moore’s supporters to admit their intentions to a human interviewer but have no such hesitation when the interview is conducted by a robot.  If so, we might attribute some of the difference between cell phone and landline results to use of human interviewers in polls that include cell users.


How Low Can He Go?

In the preceding two articles I analysed the prospects that the Democrats could regain control of the House of Representatives.  The combined effects of gerrymandering and partisan self-selection geographically mean the Democrats face a substantial burden to regain a majority in the House.  Only if the Democrats can win at least 53 percent of the national popular vote for Congress do have they have a decent chance to attain a majority.

The success of the Democrats at the polls in mid-term years depends primarily on two things, whether their party holds the White House, and the President’s job approval rating.  For the Democrats to reach that 53 percent floor needed to retake the House, President Trump’s job approval rating would have to fall to 32 percent or less.

While there have been occasional polls which put Trump’s approval in the low thirties, they have generally been outliers.  I have taken all the polling data available at Huffington Post Pollster since the Inauguration and estimated simple trends that look like this:

The bottom line represents the trend for all adults interviewed in person.  This is the sample used by Gallup, the most prolific pollster in this group.  Registered voters, or citizens interviewed over the Internet, show a slight pro-Trump bias of about one and a half points.  The top line shows the estimate for registered voters interviewed via the Internet.  (Polls of so-called “likely voters” were not significantly different from all polls of registered voters.)

If we take the plunge and extrapolate these trends to the time of next year’s mid-term election (about 650 days in office), Trump’s job approval rating falls only two more percentage points, ranging from 34.6 to 37.9 percent depending on which trend is chosen.  That is still short of the 32 percent threshold I estimated in the preceding two articles.

The results for registered voters should give the Democrats more pause.  The citizens eligible to vote have a more positive view of the President than those who remain unregistered.


Trump’s Job Approval Rating Key to Democratic Victory in 2018

In the previous article I showed that Democrats must win at least 53 percent of the national two-party vote for Congress in order to retake control of the House of Representatives.  That higher hurdle to success reflects the combined effects of more extensive partisan gerrymandering by Republican state governments and the tendency of Democrats to live in densely-populated urban districts.  These factors make Democratic votes for the House less “efficient” than Republican votes when it comes to determining which party controls the chamber.

So what combination of political and economic factors might result in a Democratic vote of 53 percent?  Political scientists have presented a number of models for mid-term elections over the years.  In an early paper, Edward Tufte showed that presidential approval and short-term changes in personal economic conditions both influenced support for the incumbent using the small sample of mid-term elections he had available at the time.  I find little support for an economic effect, but presidential job approval does play an important role.

I have analyzed both all Congressional elections and off-year elections separately.  The overall results are quite similar.  I am basing the conclusions below on the data for the seventeen off-year elections in my sample from 1950 to 2014.  Rather than treat the parties symmetrically and examine support for the President’s party as I did for the Senate, I am focused this time specifically on factors influencing support for the Democrats in off-year elections since their vote is what matters to this analysis.  It turns out just three variables account for over 90 percent of the variance in the Democratic vote for the House:

As always, the dependent variable is measured as a logit. Values above zero are associated with probabilities above 0.5; negative values represent probabilities below 0.5.  So the positive constant term indicates that the Democrats had an advantage over the period, but the coefficient for the dummy variable representing elections after 1992 is about equal in size and opposite in value.  That pattern corresponds to what we saw in the last article where Democrats had a seat advantage in the House until 1994 that vanished for two decades and has now turned significantly negative.

The other two variables capture the “referendum” aspect of off-year elections.  The Democrats do worse on average when one of their partisans occupies the White House.  However rising job approval ratings do translate into more support at the polls in the off year.  (The approval variable is coded positively for Democrats and negatively for Republicans.  If separate terms are included for Democratic and Republican presidents, the estimated coefficients are nearly identical in size but opposite in sign.  The coding I used imposes the constraint that changes in Presidential approval ratings have the same sized effect for both parties. The job approval data comes from Gallup and is based on averages of their polls near the election.)

I tried a variety of measures of economic conditions, specifically changes in real per capita disposable personal income, and none of them showed any additional effect.  I included a test of the “myopic” voter theory using only the change in income comparing the third and second quarters of the election year.  That fared no better than an approach with a longer time horizon, the growth rate over the past twelve months.  Thus there is no term in my model for economic conditions.

Since we have a Republican president, my estimates are based on the sum of the constant term and the term for elections after 1992.  If I plot the model’s predictions against President Trump’s potential approval ratings, I get this relationship:

If the President’s job approval rating falls below 32 percent, the model predicts the Democrats would win the 53.2 percent of the national House vote that we saw in the last article is required to obtain a majority of the seats in the chamber.  The last three Gallup polls reported Trump’s job approval at 38 or 39 percent.

An approval rating below thirty is historically very unlikely.  Richard Nixon in 1974 and George W. Bush in 2008 had ratings in the mid-twenties.  Jimmy Carter in 1978, George H. W. Bush in 1992, and his son in 2006 received job approval scores in the mid-thirties.  Of course, all of these incumbents had much higher ratings when they took office than did Donald Trump.

The average decline in Presidential job approval between Inauguration Day and the first subsequent off-year election has been a bit under nine points.  That would take Trump’s score down toward the mid-thirties.  However because he started at just 45 percent approval when inaugurated, he may not experience the same decline as did presidents who started from a higher rating.  For instance, it seems unlikely that Trump will experience a decline on the order of 23 points like Barack Obama did going into the 2010 midterm.   In fact, the table suggests the public treats Republican and Democratic presidents quite differently.  The Democrats all posted double-digit declines in job approval by the first mid-term election; none of the Republicans lost more than nine points over the same period, and approval for both Bush presidencies actually increased.



Can the Democrats Retake the House in 2018?

Now that all the gnashing of teeth has ended after the Republicans managed to hold on to the Georgia Sixth, perhaps we can step back and take a more systematic look at the Democrats’ prospects in 2018. Democrats will likely not make any gains in the Senate since the Republicans have only eight seats at-risk compared to twenty-three Democrats and both independents, Maine’s Angus King and Vermont’s Bernie Sanders.  That leaves the House as the only target.

There are two steps involved in answering this question.  The first is to use our historical experience with House elections to examine how votes are translated into seats.  With that information we can estimate the proportion of the two-party House vote that the Democrats need to win to take back the House in 2018.

As I wrote back in 2012, a combination of geographic clustering by party and good old partisan gerrymandering has created a “Republican bulwark” in the House since the last redistricting after the 2010 Census.  That means that the Democrats will need to win more than a majority of the popular vote for Congress if they intend to win a majority of House seats.

I have refined this simple seats and votes model in two ways.  First, I let the “swing ratio” vary between two historical periods, 1940-1992 and 1994-2016. Empirically the effects of voting “swings” on seat “swings” is significantly smaller in the more recent period.  As Tufte argues in his classic paper on the seats/votes relationship, a decline in the swing ratio indicates an increase in the proportion of “safe” seats.  As fewer and fewer seats have vote shares around fifty percent, there are consequently fewer that can be “flipped” by an equivalent shift in voters’ preferences.

I also use the results for the 2014 and 2016 elections to more sharply estimate the effect since 2010.  If we calculate the popular vote share required for the Democrats to win half the seats in the House, they would need to secure a bit over 53 percent of the (two-party) votes cast.

That brings us to the second question, what are the chances that the Democrats could win 53 percent of the Congressional vote in 2018?  Answering that question deserves an article unto itself.


It Don’t Mean a Thing …

Here are my final tests for who is ahead in the swing states.  The situation looks rather bleak for Donald Trump.


As in 2012, I am using a “chi-squared” test* to determine whether each candidate has led in so many polls in each state that it is statistically unlikely that person is not actually ahead there.  I’ve used all state polls archived at Huffington Post Pollster since June 1st and conducted a separate test using only polls conducted after the release of the “Access Hollywood” tape on October 7th where Trump claims to have committed sexual assault.  In this more recent set of polls, Arizona moves from Trump’s column to a toss-up.

In three other states, Iowa, Nevada, and Ohio, the race appears statistically tied.  Neither candidate has led in a sufficient number of polls to determine whether one of them is truly in the lead.  Hillary Clinton has a significant lead in the remaining eight states, with a total of 116 Electoral Votes.  Combined with the other solidly Democratic states, she should win at least 317 Electoral Votes on Tuesday, and as many as 347 were she to take all three of Nevada, Iowa, and Ohio.  More likely, given the data above, she will lose Iowa and Ohio and end up with 323 Electoral Votes adding Nevada to her column.


*Values of chi-squared greater than 3.84 are “significant at the 0.05 level” (with one “degree of freedom”), meaning there is a 95 percent probability that Clinton is ahead.  Values greater than 6.64 are significant at the 99 percent level.  In all eight states where Clinton has led in the polls since June 1st, her chances of actually being ahead in those states are very much higher than 99 percent. (Return)

The State of the Race

Donald Trump has gained ground over Hillary Clinton during the campaign, but the combined effect of events leaves her with a predicted five-point advantage on Election Day.

Back in 2012 I modelled the dynamics of national Presidential polling using a combination of time trends, survey methodologies, and campaign events.  In this posting I will present a similar model for the 2016 campaign using the 190 polls archived at Huffington Post Pollster covering the period from June 1st through October 25th.  All these polls include both minor party candidates, Gary Johnson and Jill Stein, in the list of alternatives.

As before I am using three types of explanatory factors to model polling dynamics:

  • a simple linear time trend that measures the number of days remaining in the campaign until Election Day; using higher-order polynomials like quadratics or cubics does not improve explained variance;
  • “dummy” variables that correspond to various features of each survey like the sample drawn (registered versus “likely” voters), the method of polling (live interviewers, automated interviewing, or via the Internet), and the identity of the polling organization;
  • dummy variables to represent various events during the course of the campaign.

For the polling organizations, I included dummies only for those who had contributed at least nine polls, or five percent of the sample.  Only six organizations met this criterion.  For the events, I included both parties’ national conventions and the first Presidential debate on September 26th.  I also included a term for the release of the “Access Hollywood” tape where Donald Trump was recorded as claiming to have engaged in sexual assault.  Because the second debate followed only two days after the release of the tape on October 9th I have combined those events together into a single dummy variable.  I have included a third variable which represents the period since the third debate on October 19th.  All dates are measured from the midpoint of each poll’s fieldwork.

Measuring the effects of the conventions was especially difficult this year since the DNC took place in the week following the RNC.  The RNC dummy is coded one starting on the close of the convention, July 21st, and extends through the following Sunday.  Eight polls were conducted during this period.  Rather than measure a separate effect for the Democratic convention, I have instead used a “post-convention” variable  that is coded as one from the close of the DNC until the first debate.  All models are estimated using “weighted least squares” with the weights proportional to the square root of each poll’s sample size.

Dependent Variable: Clinton lead over Trump
Weighted Least Squares; N=190


I present three different specifications of the model.  The first uses only the trend, method, and event variables.  The second version includes effects for the six pollsters who met the criterion of nine or more polls.  The last specification removes terms that were not statistically significant in prior specifications.  (The marginally significant effect for Ipsos/Reuters disappears in a more restricted specification.)

Starting first with the time trend, the positive value indicates that Clinton held a larger lead early in the campaign season.  A value of 0.07 means that Trump picks up about one percentage point on his opponent every fourteen days (=1/0.07).  This is a much faster pace than in 2012.  Four years ago, it took President Obama about forty-seven days to gain a single percentage point over Mitt Romney.  The constant indicates the predicted margin between the candidates on Election Day when the “Days Before” variable is zero.  Without any intervening events the model predicts a Trump victory by five to six percent.

Rather surprisingly none of the methodological variables have any effect in 2016.  Poll watchers generally expect to see a one- to two-point tilt in the Republican direction when samples are constrained to “likely” voters.  That difference reflects the generally higher propensity of Republicans to turn out since their age and social characteristics correlate with voting.  This year we see no such effect.  Nor is this likely to be a statistical artifact; polls of likely voters represent only 58 percent of the sample so there are sufficient numbers of each type of poll to generate reliable results.

In 2012, polls conducted on the Internet were about one percentage point more favorable to Obama than polls conducted by other means.  This year we see no differences between Internet polls and those conducted by live interviewers.  Two organizations, the Republican-leaning Rasmussen Reports and the Democratic-leaning Public Policy Polling, use automated calling systems where respondents are asked to enter their answers by pressing the phone’s dialpad or speaking directly to the calling robot.  Because there are only two such agencies, I included dummy variables for each of them rather than a single variable denoting the method they use.  The results for the two organizations are quite different.  Rasmussen continues to show a significant bias in favor of the Republican candidate, while PPP shows no such bias.  This difference parallels that found for 2012, where Rasmussen’s results showed a pro-Romney bias.  Rasmussen’s polling in 2016 has an even greater Republican tilt of over four points, compared to two to three points in 2012.

What the model shows most clearly, though, is the powerful effect of campaign events on the margin between the candidates.  Clinton’s lead fell after the Republican National Convention then rebounded after the Democrats convened in Philadelphia.  The debates and the release of the Access Hollywood tape further boosted Clinton’s margin.  Since the effects of these events must be measured against the overall pro-Trump trend in the polls, I have incorporated these data into a chart.
clinton-lead-trend-2 The aftermath of the conventions brought the race back to more or less the same place it was on June 1st with Clinton holding about a seven-point lead.  Her advantage decayed over the weeks that followed until the combined effects of the first debate and the release of the Access Hollywood tape again brought her lead up to nearly eight points.  The model predicts that her advantage will have fallen back to about five points on Election Day itself.  Since the model has a standard error of about 0.5 percent, the confidence interval on the Election Day prediction is roughly four to six percent.

A few other observations from these results.  First, the notion that there is a hidden vote for Donald Trump that does not appear in public polling is contradicted by the lack of any effects by polling method.  Back in January I found that Trump did over four points better in polls of Republican primary voters when they were interviewed by automated methods.  I attributed that result to the so-called “social desirability” effect; Trump supporters might have felt more shy about admitting their preference to a human interviewer.  I see no such effect in the general election polls now that Trump has been legitimated by being the Republican nominee.

Second, though I do not show the results here, including the size of the vote for the two minor-party candidates, or the proportion of undecideds, has no systematic effect on the margin between the major-party candidates.  If prospective supporters for one major candidate were disproportionately likely to defect to one of the minor candidates, or to remain undecided, we would expect to see fluctuations in the size of those groups influence the size of Clinton’s lead over Trump.  Instead it appears that potential supporters of both those candidates have moved in and out of the minor-party columns or remained undecided at roughly equal rates.  If so, as the minor candidates get squeezed as Election Day draws near, and the number of undecided voters dwindles, we should not expect to see those changes affect the competitive positions of Clinton or Trump.