Where’s the Bounce?
The big question on a lot of people’s minds is will we see any convention bounces now that both the DNC and RNC are over. My guess is no. The opinion of most voters seem to be locked in at this point. It is unusual for there to be little room for movement at this point, but that is a consequence not only of today’s partisanship, but the strong and widespread dislike and distrust of the president. If I had to guess at where we might see some movement after the conventions, I would guess that Trump would lose some support. Not a bounce, but a dip. We won’t know that until at least next week.
What we do know – and one reason it is easy for me today to make a guess that there will be no convention bounces – is that we (so far) do not see a bounce for Biden one week after the end of the Democratic convention. But, no movement is just fine for the former vice president since the polling this week shows him up by nine points (actually, 9.1) in the FiveThirtyEight polling average this morning. On August 14th, the first day of the Democratic convention, the average was 8.1. One could say that Biden did receive a bounce, but this is not a lot of movement, gets Biden back to where he was at the beginning of July, and historically “bounces” mean a temporary increase in support of about six points (which has been decreasing since 2000).
Is one point a bounce? Sure, it could be, but it is so small that it could just be statistical noise. On June 30th, Biden’s lead was 9.1 points – the same as it is this morning. During the intervening two months, his lead as gone up to 9.6 points and down to a low of 7.6 points. The low point coincided with a relative quiet period for the A-rated pollsters, so the pattern over the past two months – in the end – is steady for Biden. And that’s what he needs since he is winning over 50% of the vote in most polls and leading by anywhere from 8 to 14 points in most polls released over the past two weeks. Bounces come down, so it doesn’t really matter whether he gets one or not. Trump is the one who needs a bounce because he is so far behind. If he can get a good bounce from the convention he can try to hang on to it. But, with 50% or more of voters telling pollsters that they will not under any circumstances vote for Trump it is not clear where the bounce would come from. One place could be the undecideds, who are small this year but we see up to ten points in some polls of undecideds plus third party voters.
What does the polling mean this week? That Biden’s support and lead remains consistent and stable. This is the one key indicator this year that Biden is a stronger candidate in 2020 then Clinton was in 2016. Biden is in far better shape have a strong, but stable lead. Among really good pollsters, he is leading Trump by double-digits. It’s when the mediocre (and terrible) pollsters are incorporated into the averages that the margin goes down to nine points. FiveThirtyEight has adjustments to account for quality of pollster, but it also weighs recency favorably. So, for instance, when Rasmussen issues a poll that shows Biden leading by just four points (as it did this week) that does have a downward impact on the overall polling average. It is also almost certainly a bad poll.
Do bounces matter? Not really. They can help kickstart a campaign for the fall – like Clinton’s 30 point bounce did in 1992 (that bounce was also related to Perot quitting the campaign in the middle of the summer that year). Some bounces are smaller, but last longer. Almost all bounces are temporary – in fact, that is why it’s called a bounce. What goes up, must come down. Bounces are usually related to persuadable voters being influenced by the party’s pep rally. There was a time when the party not having the convention would stay quiet for the duration, but that’s one more norm that Trump has trampled on. So, now even your opponent’s convention week is fair game for attacks. This norm-busting also will have an impact on whether we see bounces since many voters who helped a candidate bounce may have supported both candidates during and right after the convention because they spent four days watching carefully choreographed events that put their candidate in the best possible light. For persuadable voters, these kinds of events can have a temporary effect. But, it’s mostly fleeting and the bounce dissipates.
Another sign that Biden is in good shape even without a bounce this week is that this week’s pollsters are all B, B/C, and C-rated pollsters. And even so, Biden’s lead went up this week. Next week, we may see some A-rated pollsters looking to see if either party gained an advantage (even if temporary) once both conventions are over. You can find specific polling results here.
Battleground States
From this week’s polls it looks like not only Texas is close, but so is Ohio. TargetSmart (B/C) has Biden up by one point in Ohio (47-46), and Public Policy Polling (B) has Biden also up by one point in Texas (48-47). PPP has Biden up by four points in Florida (48-44).
Biden looks good in North Carolina. Two polls have him up in the Tar Heel State. Morning Consult (B/C) has Biden up by three points (49-46), while Change Research (C-) has Biden up by just one point (48-47). Change Research also surveyed Wisconsin, finding Biden with a five point lead in the Badger State (49-44). But, the Trafalgar Group (C-) has Trump leading Wisconsin by one point (46-45). Trafalgar also has Trump up by one point in Michigan (47-46), while Change Research has Biden up by six (50-44). CR’s findings are more consistent with other recent polling in Michigan.
Perhaps the most interesting news out of the state-level polling is again in Pennsylvania. Last week, one of the nation’s best pollsters (Muhlenberg College) released a poll showing Biden up by four points. That, along with some other recent polls showing similar results, raised a red flag that all-important Keystone State might be coming off the board as a Biden state soon. This week two polls both show Biden up. One by seven points, the other by three. Change Research has the race Biden 49, Trump 46. Franklin & Marshall College (B/C) finds it Biden 49, Trump 42. The race in Pennsylvania could get close, but this week’s polling doesn’t help. One pollster shows the race falling back to the larger margin Biden previously enjoyed, but the other shows the race tightening even further. We need to keep an eye on the polling in this state.
Senate Polling
We can be skeptical of Trafalgar’s presidential results in Michigan a little easier when it also reports that Sen. Gary Peters (D) is losing. It is possible that this C minus rated pollster has cracked the code in Michigan where other pollsters haven’t. But, until we see more data from other pollsters, Trafalgar’s findings this week in the Wolverine State have to be considered outliers. Here’s a run-down of the rest of this week’s Senate polls.
Colorado: Public Policy Polling (B): Hickenlooper 51%, Gardner 42%; Hickenlooper +9
Michigan: Trafalgar Group (C-): Peters (D) 47%, James (R) 48%; James +2
Michigan: Change Research (C-): Peters (D) 50%, James 45% (R); Peters +5
North Carolina: Change Research (C-): Cunningham (D) 52%, Tillis (R) 42%; Cunningham +10
North Carolina: Morning Consult (B/C): Cunningham (D) 47%, Tillis (R) 39%; Cunningham +8
Virginia: Roanoke College (B/C): Warner (D) 55%, Gade 34%; Warner +21
At least three polls came out in Massachusetts for the Democratic primary between Sen. Ed Markey and Rep. Joe Kennedy III, but FiveThirtyEight has not reported on them yet. Markey is winning convincingly in each poll. Data for Progress (B-) has Markey up by seven points (50-43). UMass Lowell/YouGov (B) has the incumbent with the largest margin (12 points) in a 52-40 lead; Suffolk University (A) has Markey winning by 10 points (51-41). The race appears to have significantly turned against Kennedy, who was the early favorite in this race, over the summer. The Massachusetts primary is this coming Tuesday, September 1st.
Labor Day is just around the corner and we are less than ten weeks to November 3rd. Get ready for a busy fall!