It might be the final poll of the Louisiana governor’s race, and it’s the best result for Eddie Rispone supporters of the lot. But whether pollster John Couvillon’s survey of 600 likely voters across the state released Thursday accurately reflects what’s going on among the electorate is something we won’t know until Saturday night.
Late Saturday night, from the looks of it.
So here are five thoughts about the poll and what it portends, assuming it’s correct.
1. The sample is realistic.
What we can say is that Couvillon probably has a pretty good bead on what the electorate will look like on Saturday. He pegged the African-American share of the vote at 29.5 percent, which is higher than the 27.6 percent it was in the primary but lower than the 31.1 percent it was during early voting for Saturday’s runoff. Couvillon’s 29.5 percent black share of the vote is also lower than the 31 percent Monday’s Mason Dixon poll assigned.
Couvillon also assigned a 45-35-20 Democrat-to-Republican-to-Independent ratio of the sample by party. That’s a number which from a conservative perspective is, well, conservative. The D-R-I ratio in the primary was 44.9-37.1-18.0, and in early voting for Saturday’s runoff it was 46.4-38.1-15.5. What to take from this is there are probably going to be more Republicans in the electorate than the 35 percent Couvillon has laid in for his sample, and fewer independents. He’s got the Democrats more or less right.
Our guess is that the black vote will be closer to 28.5 than 29.5, and probably not 31 percent. We’ll elaborate on that below.
2. It’s very close to the Mason Dixon poll.
The numbers in the JMC Analytics poll if you adjust the sample, come out almost identical to that Mason Dixon poll on Monday. Mason Dixon had it 48-46 Edwards, but that was with a 31 percent black vote. Move that number down to 29.5 and it’s entirely likely Edwards drops to 46 or so. The JMC Analytics survey has him at 45 percent, with Rispone at 45.5.
The point being that polling of the race is starting to crystallize a little, and the differences in the various poll results are starting to be driven by differences in the sample. This is typical of the late stages of a race and it’s a reason why polling close to Election Day is often fairly accurate.
Both polls show the race practically tied. More importantly…
3. Edwards is under 50 less than a week before the election.
This is a big deal, because it illuminates something political pros will tell you over and over – which is that any incumbent who can’t consistently poll above 50 percent is going to have a real problem winning.
Moreover, Monday’s Mason Dixon poll had Edwards at 48 with black voters, whom he’s going to get a good 90 percent of if not more, at 31 percent of the vote. Couvillon, with a sample we think is closer to reality, had him at 45.
Couvillon’s poll was conducted Tuesday and Wednesday, as opposed to the Mason Dixon poll which was conducted Nov. 5-7 (Tuesday to Thursday of last week). That’s an indication the direction of Edwards’ support might be dropping, rather than increasing.
You can’t be an incumbent at 45 percent with two days to Election Day and win a race. If this poll and the Mason Dixon poll are correct, and the electorate looks like Couvillon’s theory or even friendlier to Rispone, and Edwards is cooked.
4. The hype about early voting was probably off-base.
This is less about the JMC poll itself and more a defense of its assumptions, because there will be people who’ll say the 29.5 percent black share he’s assigning is too low.
That 31.1 percent number Edwards’ camp was able to rack up in early voting is a function of what seems to be a pretty successful “Souls to the Polls” campaign they put together. Let’s call it very successful, because if that 31.1 percent number were to hold up it would break the modern record for black voting as a share of the Louisiana electorate in a statewide race. The current record is 30.6 percent, which is what Barack Obama pulled off in 2008, his first presidential run.
If you think John Bel Edwards is a bigger draw for the black vote than Barack Obama in Louisiana, then…well, okay.
There is something else, and we somewhat covered this in our post earlier in the week about polls and the early voting. Edwards’ turnout guru Scott Arceneaux, an old Louisiana Democrat political hand who’s been working in Florida the last few years, is credited with the big early-voting push last week. But Arceneaux has a history of big early voting numbers, particularly in Florida for Hillary Clinton in 2016 and Andrew Gillum in the 2018 gubernatorial race, only to watch the promise of those numbers go up in smoke on Election Night. Clinton wasn’t supposed to lose to Donald Trump, but she did, and the hype about Gillum’s incredible early vote numbers completely evaporated when Ron DeSantis won a relatively easy race for governor of that state two years later.
One reason why has to do with the different dynamics of trying to do vote-hauling in early voting, as opposed to on Election Day. If you’re busing people to the polls, it’s a whole lot easier getting people to assemble in a church parking lot somewhere, let’s say, and everybody goes to one, or one of very few, voting places, which is how it works with early voting. On Election Day, you can’t collect huge numbers of people at one place, feed them jambalaya or fried chicken or ribs or whatever and then put them on buses, because instead of one big voting place you’ve now got 50 precincts they need to go to. It’s a lot bigger logistical challenge.
So the question isn’t really whether the 31 percent number can hold up. It’s more how far it drops. Anything below 29.5 percent and Edwards really can’t win.
5. Nothing whatsoever is decided, and it is literally all about turnout.
Rispone had a pretty good day Thursday, given that the word was there were 40,000 people who had requested tickets to President Trump’s rally in Bossier City, which was held in an arena which only held 14,000, and meanwhile Edwards showed up in Shreveport earlier in the day for something which was considerably less raucous.
Edwards spoke to reporters at the AFL-CIO’s office in Shreveport about seven hours before Trump rallies conservative voters for the third time in a month to defeat Edwards, a Democrat. The rally will take place at the CenturyLink Center in Bossier City.
A dozen people manning a phone bank at folding tables stopped making their calls to listen to the governor in Shreveport.
One thing the JMC poll does indicate, that we’ve heard a lot of anecdotal evidence of, is that the voters aren’t overly enthralled by either campaign. You’re supposed to have a smaller number of undecideds, not a larger number, at the end of a race, and yet Couvillon found 9.5 percent undecided. The Mason Dixon poll taken a week earlier had six percent.
Undecided voters this late in a race, regardless of what they might tell a pollster, are people who probably don’t show up to vote.
But some of them will. Who they vote for likely depends on friends, neighbors and other influencers. So if you care about the outcome of this race, you’re going to want to do some electioneering of your own, and make sure it’s not just you getting out to vote – you’re going to want to get your people to the polls.
That’s the game when the race is too close to call, which is what this one is.