Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

Convention bounces in a polarized era

July 22nd, 2016, 9:55am by Sam Wang

Measuring the effect of a particular political event is challenging. Any single poll spans multiple days, and multiple polls are necessary to get good accuracy. National surveys give the first indication, within a week. State polls (upon which the PEC snapshot and forecast are based) are more accurate when aggregated, but take longer.

In the case of the Republican convention, we will have a hard time knowing what its effects are in isolation. Certainly the event was distinctive. Last night, Donald Trump entered the general election campaign with a harsh 76-minute speech that painted the U.S. as a dystopia, and his opponent as a criminal. These claims do not hold up to scrutiny – but they do show his approach for the months ahead. One might like to know the net effect of that speech, that of his endorsers and various other Trumps, and of Ted Cruz, who called for citizens to “vote your conscience.” It would be interesting to know if such a convention would close his deficit with Republican voters.

Almost immediately, the Democrats now take the stage. Today, Hillary Clinton is expected to announce her vice-presidential pick. Next week comes her party’s convention. Many polling measurements will capture the combined effect of all of the Republican and Democratic events. The conventions are close in time, making it easy for viewers to see contrasts between the parties.┬áMy guess is that the net effect should be relatively large, favoring the Democrats. If I am wrong, undecideds should still decrease.

However, how large is “relative”? The effects of conventions have been declining. Based on Gallup data, the median “net impact” (more-likely-to-support minus less-likely-to-support) of conventions from 2004-2012 was 5 percentage points, compared with 16 percentage points for 1984-2000. I would characterize this as entrenchment of voters, a feature of political polarization. I will be interested to see if the net impact of this year’s conventions is an exception to the trend.

Today the electoral snapshot is at Clinton 312 EV, Trump 226 EV. Margins in many states are quite narrow – note the pastel appearance of the electoral maps in the right-hand column. Consequently the Meta-Margin (how much swing toward Trump would be needed to create an electoral tie) is only 2.5%. This close lead, in PEC’s approach to prediction, ┬ámakes Clinton’s November win probability only 80%, in a range (20-80%) that I call uncertain. This is our starting point for evaluating the weeks ahead.

Tags: 2016 Election

26 Comments so far ↓

  • Matt McIrvin

    I’ve been seeing claims of a convention bounce on the basis of movement that mostly happened before the convention, I think because people didn’t look closely enough at when the polls were in the field. I think the best you can say so far is that the first couple of days of the convention didn’t obviously cause Trump’s numbers to decline.

  • bks

    I have no specific memories of the 1992 convention (my preferred candidate that year was Jerry Brown, and he’d still be my preferred candidate if he were running). I wonder what accounts for the Dem peak that year.

    • Froggy

      The race in 1992 was very volatile because of the presence of Ross Perot, who was actually the front runner in some polls as late as May. That’s a measure of how discontented people felt with the two parties, and it provided the opportunity for a large improvement in perception.

      There was also over a month between the end of the Democratic convention in July, and the Republican convention in August. The Democrats took advantage of that by having Clinton and Gore embark on a 1,000+ mile bus tour immediately following the convention. This was a novel thing at the time, and it sustained the attention and good feelings of the convention much longer than usual.

    • Froggy

      Reading up again on the 1992 campaign this afternoon, I see that Ross Perot announced on the last day of the Democratic convention that he was withdrawing from the race. (He re-entered at the beginning of October.)

      So a major rival withdrawing during a convention produced a huge convention bounce. But such an event is unlikely to recur.

  • Josh

    What’s interesting to me is that we appear to have a situation where national polling has been very, for lack of a better term, inelastic, with Clinton basically staying between 4 and 6 points ahead of Trump since the primaries effectively ended in March/April. This inelasticity isn’t substantially different from 2012, when Obama generally led Romney by between 1 and 3 points nationally for most of the general campaign.

    The main difference this year is that there also seem to be an abnormally large number of undecided voters. It’s not a situation like in 2012, where part of the reason polling didn’t fluctuate much was because almost everybody (90%+) of voters had made up their mind by early summer. I wonder if there will be a blitz where this group of undecideds makes up their mind seemingly all at once, or if the effect will trickle in gradually over the next 2-3 months.

    Is there a precedent for this kind of situation, where the electorate is highly polarized and the polls aren’t moving much, but there are simultaneously a large number of undecideds?

    • Joseph

      Is there any information about the makeup of the undecided vote? If it’s split evenly between Repubs and Dems, that’s one thing. If it’s mostly Repubs, that could work against Ms Clinton.

    • Michael Coppola


      A number of the national polls have breakdowns by party affiliation. You can check them out at HuffPollster. Looks like undecideds are more independents than either Rs or Ds, as you might expect.

  • 538 Refugee

    Oh, I thought you said “slur”. ;)

  • DaveM

    I’ve been inoculating myself against the media-fostered impression of wild national polling fluctuations by applying the median-of-all-polls-whose-end-date-is-within-the-last-seven-days calculation, and it’s been remarkably stable at Clinton+3 for quite a number of days now.

    It’s hard to know whether this reflects a relatively static electorate, or simultaneous, mutually offsetting gains and losses on the Trump side, as he consolidates one cohort of his potential supporters while driving off another…

  • 538 Refugee

    Trump Would Fund Super-PACs Aimed at Taking Down Cruz, Kasich

    This guy in charge of the military? A guy that is this vindictive? Let’s think about this. Putin and Kim Jong-Un have both endorsed Trump. What would a ‘military coup’ look like in the US? Would they simply refuse to follow orders? I’m going to have to look up and see what powers the president has in appointing loyalists as military leaders.

    • Matt McIrvin

      Which ought to be a red flag in itself. (Gravis’s track record is not great; more generally, getting worked up over individual polls is a bad idea.)

    • Josh

      Several problems here….

      1) Gravis only calls landlines

      2) Gravis only does robocalls

      3) Gravis has a well-documented R lean

      But the biggest problem is that this is a single poll in July, making it worth, on its own, close to nothing in terms of describing the current state of the race.

    • Michael Coppola

      So was this a robopoll with only one question and only two choices? Wouldn’t count undecideds by definition, since those people simply wouldn’t answer the question.

  • Matt McIrvin

    Boy, 538’s model is getting really weird–they’re claiming Trump would win some states today on the basis of polls that almost entirely show Clinton in the lead, because they’re old polls and they’re doing some kind of extrapolative trendline adjustment. They might be too clever for their own good.

    • Slartibartfast

      I was looking at North Carolina in 538’s model a few days ago and their 5th most heavily weighted poll was a Survey USA poll from March which had Trump +7 (rejiggered to Trump +10).

      Nate and company obviously have reasons for all of the things they do, but the more I see how many modeling decisions requiring justification are in 538’s model (even the polls-only one), the more I appreciate the simplicity and quality of the PEC’s methodology.

      As professor Wang has documented (+5.5% undecideds), this is an unusual year and the more bells and whistles in an analysis, the more likely an assumption that might have been true in the past (and the future), but isn’t correct in this environment has been made.

      The only thing I believe about the polls right now is Sam’s comment that they are at their least accurate.

    • Phoenix Woman

      I stopped reading 538 once I realized how much they fudged things in order to keep the clicks coming. Contrast that with Prof. Wang, who actually took a couple of weeks off back in the spring once it was clear that Hillary had the nomination locked up. He doesn’t depend on this site as his sole means of support, so he doesn’t feel the pressure to fudge the numbers just to get more site traffic.

    • Bill G.

      @Matt McIrvin: Part of their model weights by past results, I think you’re seeing that play out a lot in states like North Carolina right now. I used to be more of a believer in the weighting (of many kinds not just historical), but Sam Wang here has me pretty convinced that it’s not necessary, even to account for polling quality differences between firms.

      @Phoenix Woman: I try no to throw around accusations like that, but as a long time 538 reader I’m finding it hard not to agree with you. There has definitely been a gradual decline towards, something, click grabbing I guess over time there. You didn’t see it much at all when they were independent (could be rose tinted glasses, though for all I know), and really it wasn’t too bad four years ago when they were at the NYT (probably because it’s still an actual source of journalism), but ever since being bought by ESPN/ABC/Disney, well, I think it’s hard to deny that there’s a drive to keep people coming back. I don’t think that they do anything terribly untoward, such as manipulating data, but it definitely behooves them to write things that keep people coming back. I could honestly do without the constant editorials there, and I think that is what has gotten them in trouble this election cycle. I miss when it was just Nate Silver, and occasionally a guest, writing about what specifically the polls showed, kind of like here really.

  • Amitabh Lath

    Here’s something to watch for that might move the polls: Josh Marshall has a post on the Trump-Putin connection.

    Apparently a lot of Trump Inc. income comes from Russian oligarchs (who by now are all in Putin’s orbit) and coincidentally Trump is making overt (telling the NYTimes that the US might not come to the aid of NATO allies in Eastern Europe) and covert (removing the plank in the GOP platform that called for aide to Ukraine in case of Russian attack).

    Of course this could all fizzle out, but even if there is no smoking gun it does bring together various storylines of murky finances and love of Putin and reduction of our military footprint in Europe. It may do to Trump what the “47%” did to Romney, reinforce an emerging sterotype.

    • 538 Refugee

      I’ve read Jr.’s statements about the Russian money before. Probably just one more reason we will never see his tax returns. Since Trump was forced to forgive the $50 million campaign loan to get others to contribute I wonder if a case can be made for illegal foreign money being funneled into the campaign? At least in a political ad sense? Calls for hearings and renewed pressure to release his returns should ramp up at some point.

      I remember reading the Republicans used about 20% of what the Democrats had on Trump in the primaries. The Republicans didn’t do their homework thinking he would just fade away.

    • Matt McIrvin

      What I’m also interested in here is to what extent Putin might be trying to fan the embers of the Bernie-or-Bust phenomenon and split the Democratic coalition, by selectively dumping data through Wikileaks, bashing Democratic foreign policy from a “left” perspective on, etc.

  • jharp

    Gallup called me today asking questions about Trump and the Republican Convention.

    No evidence of any bounce from my answers.

  • 538 Refugee

    We have been told that ‘in private’ Trump is a much different person that the one we see publicly. Not so according to the guy that followed him around for 18 months so he could write “The Art of the Deal”. It turns out Donald didn’t even have the attention span to sit around on Saturday mornings to be interviewed for the book.

    Years ago I read a book. I believe “My Lord of Canterbury”? I seem to recall at the end the priest, as he is being burned at the stake, holds his hand out into the flame as if to burn away the sins it had created. This pretty much sums up the ghostwriters feelings about his contribution to the rise of Trump.

  • A New Jersey Farmer

    New Georgia poll has Clinton within one. That’s got to alarm the righties.

    • 538 Refugee

      I’m kinda waiting for some quality state polling to get rolling. I hope it’s soon. I think it was Clinton’s campaign manager that said last night they don’t pay any attention to public polling. I’d love to see what they’re looking at and what their analysis of it is.

Leave a Comment