Primary Election Polling and Turnout

This post is about pre-election polling and turnout expectations; I’ll wait until all the numbers from the Secretary of State are official before I go into any more detail.

Primary Election Polling

Here’s the final GPI update I posted before the primary election:



And here are the final results (unofficial):



The polling for Mark Dayton and Matt Entenza was pretty accurate; the polling for Margaret Anderson Kelliher was way off. Why was this? For one thing primary elections are notoriously difficult to poll, not only because of questions about who’s going to actually show up on Election Day, but also because in an intra party election, voter preference is likely to be more fluid than in a general election, in a three way race that tendency is even more pronounced. This election also featured the added uncertainty of a new primary election date in the summer to further compound the pollster’s difficulties.

But how do you get two candidates so close and miss one so badly? What the GPI doesn’t show is the 9.33% undecided number that, if added to MAK’s total, puts her much closer to where she finished. Of course you can’t assume that 100% of undecided voters broke for MAK, in fact, my assumption before the primary was that undecided voters would break her way at about a 70-30 split with Dayton, weather that actually happened, who knows.

Clearly MAK had the superior ground game, but that appears to have been largely confined to the metro area where she was able to turn out her base. Outside of the metro area was a different story and just as he did in the 2000 primary Mark Dayton cleaned up in the non-metro counties. The difference, as many suspected, turned out to be St. Louis County, where thanks to the help of Yvonne Prettner Solon, Mark Dayton did better than probably anyone expected.

Again, these numbers are unofficial (and incomplete as St. Louis County isn’t fully reported as of this writing) but Dayton beat MAK 56-29, if she could have kept that margin respectable she could have won, but to do that badly in such a key county is the main reason she lost. Joe predicted before the election that Prettner Solon would be the x-factor and it appears he was right.

In the end I suspect the discrepancy in the polls and the results is due to two factors, undecided voters breaking for MAK (something which will be difficult to prove) and a fantastic turnout operation (something that can be looked at when we have final SOS numbers).

The other option is that she was always in contention and the polls somehow completely missed a good portion of her support. Seeing how much space I’ve given to the two theories you can probably guess which one I favor.

Turnout

Here is the expected turnout table I posted before the primary:



Unofficial turnout stands at 441,982 as of this writing, with 99.93% of precincts reporting, so we can guess that the final number will be about 442,500. This falls comfortably into the “no penalty” category of adjustments for the summer date, meaning that when using 1998 as a turnout template there was no perceivable drop off in primary participation due to the summer primary date.

The reason for this is likely twofold; a record number of absentee ballots cast in a primary election combined with the Kelliher campaign’s (and maybe to a lesser extent the other two DFL campaigns) effective turnout effort. Once we get final numbers from the SOS, which will certify results on the 17th, all of these theories can be tested to a certain degree, but until than let the speculation begin.

The party for the party is at Jax

I was liveblogging primary night from Jax Cafe, site of the DFL unity party, here is the transcript, from Minnesota Progressive Project:

[Update 12:40pm]: And I was wrong about the concession speech too, Dayton's lead just ended up being too large, in relative terms.

That's it from Jax where the drinks are too damn expensive for a DFL event. Seriously people, this needs to change next time. If I can't get a beer for a decent price why do I even want to go?



[Update 12:17pm]: As Dayton pulls past the 2,000 vote mark there is still no sign of Margaret. I doubt we'll get a concession tonight as I'm sure she'll want to wait for the re-canvassing, just in case.

[Update 12:03pm]: 92% in and Dayton now has a 1300 vote lead. I might have been premature in concluding this is heading for a recount. Dayton might win it outright.

[Update 11:52pm]: 91% in and Dayton is now winning, by only 335 votes.

[Update 11:48pm]: It's down to a 60 vote margin with 90% in. This one's going to a recount folks.

As soon as I write that Amy comes to the podium to give us an update.



[Update 11:30pm]: It's down to 425 votes with 87% in.



[Update 11:22pm]: WCCO appears to be ahead of the SOS website, they are reporting that with 86% in the margin is less than 1,300 votes.

[Update 11:10pm]: A number of the counties yet to report are counties that Dayton did well in during his 2000 primary race. These include Aitkin, Cass, Lincoln, Roseau and Swift.

Of course what I said earlier still holds, if MAK can keep it close in St. Louis County she can win. With just over 7% in though, Dayton is leading 60% to 25%.



[Update 11:01pm]: With over 65% in it's now a two point race. St. Louis County looks like it will decide the election.

[Update 10:50pm]: Turnout right now is at almost 350,000 and it looks like it could approach 400,000.

[Update 10:42pm]: 60% in and it's now MAK 41.89, Dayton 39.25 and the SOS site only shows 7% in from St. Louis County.

[Update 10:35pm]: Chris Coleman is now speaking.



[Update 10:30pm]: The race just keeps getting closer. With 52% in it's 42.33 to 38.87, only 5 points of difference now with Ramsey and Hennepin almost all reporting. It's going to be a tight one.



The Big E has left the building but below is a picture of him and me hard at work. The only thing missing is a beer.



[Update 10:20pm]: A few elections that can be called; Barb Goodwin will be the nominee in SD50. Paul "not the gay one" Gazelka will likely win in SD12, which could open up that seat for a DFL takeover. In SD67 John Harrington will be the nominee. In 15B Zachary Dorholt will get the opportunity to take on King Banaian and in 65A Rena Moran is the winner.

[Update 10:03pm]: John Gunyu is speaking right now after a blackout/internet issue caused a slight delay and bit of nervousness.



[Update 9:38pm]: With almost 30% in reporting in SD50 Barb Goodwin is winning 70-30 over Satveer Chaudhary.



[Update 9:33pm]: Al is speaking. He sure knows how to whip up a crowd.



[Update 9:32pm]: Another update; Entenza concedes.

[Update 9:21pm]: Amy is revving up the crowd right now. You can feel the excitement building with every minute that goes by.

"It's happening." Someone shoouts.




[Update 9:10pm]: With over 92% of precincts reporting it looks like John Harrington is going to win in SD67.

[Update 8:50pm]: Tony Sertich is speaking right now. He gave us the update we were all waiting for, Twins 8, White Sox 3.

[Update 8:35pm]: The latest update from the SOS website shows Hennipen county 81% in and Kelliher with 48% and Dayton with under 35%. These are good numbers for her.

[Update 8:30pm]: The place is buzzing now with red shirts everywhere. People are excited about seeing Margaret in the lead, they have the Secretary of State's website up on a projection screen for everyone to see the results coming in.

The early lead for Kelliher we're seeing is coming in from Ramsey and Wright counties primarily. She needs to keep that up to pull this one out.

[Update 8:00pm Polls are Closed!]: The room is filling up and you can feel the anxiety in the room, people with the campaign are nervous but optimistic.


[7:15pm]: I just got to Jax and plugged into the intertubes. Some of the networks are here, MPR and some other bloggers, but no one else yet.

Amy, Al and Keith are all supposed to show up, along with Margaret of course. I'll be updating this thread with all the latest news from here as things develop.



Robin (she's taking all the pictures)



Tony (I'm typing all the words)

What to watch for on Primary day

This is your 2010 primary returns watching guide.

Governor

Looking through the crosstabs of the SurveyUSA poll from last week, there aren’t really any major areas of weakness for Mark Dayton. The few areas where his support might be considered soft are young people and college grads. Looking at the regional breakdown his toplines are strong everywhere, but in the northeast he’s "only" leading Margaret Anderson Kelliher by 10 points 41-31.

I looked at the results of the 2000 Senate primary, Mark Dayton’s last statewide primary race, for any regional weaknesses but his support was surprisingly widespread. The few areas in which he did poorly were somewhat correlated with areas of strength for DFL endorsed candidate Jerry Janezich. Those areas were primarily in the northeast, the same regional area that MAK polls closest in.

If MAK is going to win though, she has to do better in the Twin Cities than the polls have shown. If she can’t get to parity in her base areas she doesn’t really have a chance. In order for her to win her grassroots focused campaign has to turn out the Twin Cities and St. Louis County, otherwise it’s Dayton in the general.

Minnesota Senate

If you have questions about how to read the hPVI listing for the individual districts here is an explanation.

SD50



This is the undercard to the Governors primary, pitting current DFL Sen. Satveer Chaudhary against former DFL Rep. Barb Goodwin in a race that will answer the question, can a DFL state senator keep his job even after pissing off virtually everyone in the party.

Living in the district I see about an equal split in yard signs, for whatever that’s worth. This one is hard to forecast but may be a better indicator than the governor’s race of the DFL party’s influence in primary elections.

SD12



GOP Sen. Paul Koering is gay. His challenger, Paul Gazelka, is not gay. Sen. Koering had dinner with a gay porn star, Mr. Gazelka did not. For these reasons the state GOP has kicked their incumbent candidate to the curb for a generic republican, who they want you to know is not gay (or at least has not come out yet).

This will be an interesting race to watch for the above reason and the possibility of a DFL pickup in November by nominee Taylor Stevenson. While SD12 is an R+14 district, its two house seats are held by DFLers, making a pickup of the senate seat possible, especially if the two GOP factions have trouble coalescing around a candidate after the primary.

SD28



Long time Senator Steve Murphy is retiring in this R+8 district, and there are two DFLers, endorsed candidate Joe Fricke and Bruce Montplaisir, running for the chance to hold the seat from GOP candidate and Red Wing Mayor John Howe.

This race is sort of the opposite of SD12’s, a hold here will be difficult without an incumbent in the race and with both house districts held by GOPers. On the other hand in this political environment maybe not having an incumbent will work in the DFLs advantage against a candidate who is an office holder.

SD67



There are nine DFLers on the ballot in this race and the winner gets a senate seat. Here’s a good primer on who some of the candidates are.

Minnesota House

15B



Two DFL candidates are vying for Larry Haws St. Cloud house seat, DFL endorsed candidate Zachary Dorholt and Carol Lewis, with the winner getting the privilege of taking on none other than King Banaian in the general election.

The two candidates are about as different as you can find, Dorholt is 29 and running in his first election while Lewis is 53, a former school board member and has been involved with St. Cloud politics for 20 years. 15B is in Tarryl Clark’s vacated senate district, so there won’t be a single DFL incumbent running for reelection to the legislature in the district, something which may or may not matter in the general.

65A



Jeremiah Ellis and Rena Moran are facing off for Cy Thao’s open seat, the general election being a mere formality in this D+56 district.

This race is worth watching primarily because Ellis is the DFL endorsed candidate and Moran is the TakeAction endorsed candidate, making this a proxy battle of sorts.

17B



Jeremy Kalin is leaving the house after only two terms and Cindy Erickson is running unopposed in the DFL primary to replace him. The action here is on the republican side as two candidates will face off, insurgent Sheldon Anderson and the GOP endorsed candidate Bob Barrett.

At R+15, this is a vulnerable seat and could be in play, but fortunately the two GOP candidates appear to be trying their best to out teabag one another. Bob Barrett won the endorsement against Sheldon Anderson on the fourth ballot, but Anderson claimed the process was unethical and decided to run in the primary even though he promised to abide by the endorsement. Good times.

Estimating Primary Turnout

Even though I’ve already submitted an answer to the Primary Prediction Contest I’m going to take a crack at actually running some numbers to come up with a more accurate turnout prediction.

As you can see from the graph below, turnout in Minnesota primaries as a percentage of eligible voters has fallen over the years.



But turnout in absolute numbers has remained basically the same. This is due to the year by year steady growth in eligible voters.



The peaks in the above graph roughly correspond to when there were competitive primaries, which there hasn’t been the last two cycles. For that reason it’s not really fair to use the turnout percentages of the two previous cycles to estimate this one, with a three person DFL race and an IP primary that will drive turnout higher.

This is from the Winona Daily News supports that conclution:

Nearly 22,000 voters have cast absentee ballots ahead of Tuesday's primary, a level of absentee turnout that already tops the last two gubernatorial elections.

If we were to look to Minnesota’s recent past to find a similar primary 1998 is the first one that sticks out, that year there was an open seat election with multiple DFL challengers. Overall turnout that year was just north of 20%, with the DFL turnout at around 12%. I’m going to use that 12% figure, actually 11.8%, as a baseline for estimating turnout going forward.

Now the question becomes how to factor in the change from a September date to an August one. What I did was to figure out the Democratic turnout, as a percentage of eligible voters, for states with non-summer primaries and for states with summer primaries. I limited this to states that have similar PVI’s to Minnesota’s (within half a standard deviation away), since I’m attempting to measure turnout in a partisan election I figured I should try and control for a states partisan tilt, and I used information from the last two gubernatorial election cycles in those states.

The three non-summer states I used are Ohio (5/4), Pennsylvania (5/18) and Oregon (5/18). The summer election states are New Mexico (6/1), Nevada (6/8) and Michigan (8/3). Why are New Mexico and Nevada considered summer primaries? Because I needed data points, that’s why. If I really wanted to be meticulous about this I could have gathered information from every state and attempted to figure out a turnout curve based on time of year the primary is held, but I didn’t want to devote nearly that much time to this problem. Instead I’m just going to use a small sample of states with similar partisan makeups.

What this analysis shows is that states with summer primaries experience an approximately %15 percent decrease in primary turnout compared with states that have non summer primaries. The three non-summer states average about 11% turnout and the three summer states average a little over 9%.

If we use 15% as our “summer penalty” the turnout estimate comes in at 381,850. There are a few things that need to be considered though, first off Oregon votes by mail which leads to higher turnout in that state than there would otherwise be. Additionally, in Pennsylvania this year the gubernatorial primary was overshadowed by the Sestak-Spector senate primary, which likely drove turnout higher.

Because of those factors and the uncertainty inherent in such a small sample (twelve elections) I’m going to use 10% as a summer penalty, instead of 15%. With a 10% penalty the turnout estimate is 404,312. I would say than that turnout is likely to be between 375,000 and 425,000. That is my actual estimate of turnout.

What if we apply no penalty though, what if the change in the election date is offset by the GOTV efforts of the three DFL campaigns, if this is the case than the turnout estimate is 449,235. We’ll see after the primary which, if any, method arrives at the right conclusion, but here are the guesstimates all together.

New SurveyUSA poll, Dayton widens lead

A new KSTP/SurveyUSA poll was released last night and it shows Mark Dayton solidifying his lead in the DFL primary to be held on Tuesday.

Unlike the previously released StarTribune poll that many people criticized for its small sample size, this poll doesn’t suffer from that issue and it confirms what pretty much every other poll of this race has shown, including that StarTribune poll; Mark Dayton has a commanding lead that will be difficult for the other candidates to overcome.

Here are the toplines with the numbers from the previous SurveyUSA poll in parenthesis.



Dayton is the only candidate who has shown any significant movement from the last poll and he continues to poll at or around 40%. There is almost no way either of the other two candidates can win with Dayton pulling over 40% unless there is a mass exodus of support from Matt Entenza to Margaret Anderson Kelliher which seems very unlikely at this point.

Here’s the updated primary GPI:



This is the biggest lead that Dayton has held in the GPI up till now.

There is also a change in the general election numbers, with Dayton pulling in front of the other DFLers against Emmer and Horner.



Kelliher is +6 against Tom Emmer, Entenza is +5, but Mark Dayton is +14, an increase of 11 points since the last SurveyUSA poll. Now most of that is probably due to Tom Emmer himself and his imploding campaign, but it appears that Dayton is also doing better than the other DFLers among Horner voters and undecided voters.

Here’s the update general election GPI:

How to read the Minnesota Legislative hPVI Chart

I've had a couple requests to further explain the Minnesota Legislative hPVI Chart I posted on Monday, so here goes.

Here is one district in isolation:



Going left to right, the first cell is blue with an 8 in it. The 8 means that this cell represents Senate district 8, the blue represents the fact that this seat is held by a DFLer. The next two cells, A and B, represent house districts 8A and 8B and the blue, again, represents that these two seats are held by DFLers. In the first three columns the color coding will either be solid blue or solid red, a district cannot be partially held by a DFLer or GOPer so there is not any shading of these cells.

Moving on to the fourth cell from the left, this is where we get into the hPVI information. The heading "SD" means that this cell represents the Senate district of the row, which is 8. The 8 in the first cell means that this entire row is referring to Senate district 8. So we can see that SD8 has an hPVI of D+6, that's because the way I've formatted the hPVI information a positive number means a Democratic lean and a negative number means a Republican lean. So a positive 6 means that SD8 is 6 points more Democratic than the State as a whole. I did it that way to make doing the math easier, not to imply that somehow a republican district is a negative thing. When you see a -10, it should be read as R+10.

This cell is also shaded slightly blue, for the hPVI information the cells are not solid blue or red, they are shaded depending on how the hPVI number relates to the others. So the cell representing the hPVI of SD8 is shaded slightly blue, because it's slightly more Democratic than average. You can see that the cell representing house district 8A is a deeper blue, this is because it has a higher hPVI and is more Democratic.

The column at the very end is the hPVI difference between the two house districts, how much of a partisan divide there is within the Senate district. SD8 has the largest divide of any Minnesota Senate district which is why it's solid green if the divide was smaller the cell would be less green and more white.

Here's another district in isolation:



So here we can see that all of the seats in district 56 are held by DFLers even though the hPVI of all the districts has a Republican tilt. And you can see in the last column that the two house districts have basically the same hPVI.

Here's another example:



Here you can see that SD57 and 57A are held by DFLers and 57B is held by a GOPer. SD57 has no partisan tilt, A is slightly Democratic and B is slightly Republican.

I hope that helps anyone who was confused.

GPI update

On Sunday the StarTribune released the numbers of their "Minnesota Poll." Here's what the Gubernatorial Polling Index looks like with the addition of that poll:



All of the DFLers now lead Tom Emmer whose campaign is crashing and burning before our very eyes. But that's not what is of primary interest right now (did you catch that sweet pun!). The DFL primary is next Tuesday and this is the first poll on that race in over a month.

Here's what the DFL primary looks like with the addition of the StarTribune poll:



Because the only other poll currently in the primary GPI, besides the StarTribune poll, is an almost 50 day old SurveyUSA poll, the numbers will look very similar to those of the Strib poll.

So what does the Strib poll tell us? It tells us that the race is in virtually the same place it was back in May. If you don't include the Mellman poll, Mark Dayton and Margaret Anderson Kelliher's numbers have been virtually identical in the three polls of the race so far and while Matt Entenza showed movement in the SurveyUSA poll that movement needed to continue for him to be competitive and it hasn't.

Again, not including the Mellman poll there hasn't been any poll released to date that's shown MAK within single digits of Dayton. This is a very important point. A good ground campaign can certainly be worth a few points, but 10 points seems like a stretch.

There has been some criticism of the primary part of this poll, mainly from the MAK and Entenza campaigns. The high margin of error is certainly something to be aware of, but it doesn't render the poll useless. Primary election polling is a more challenging endeavor than general election polling because it's anyone's guess what the primary electorate will look like. It's made more difficult this year because DFLers are not used to voting in early August. So no one really knows who will show up on Election Day.

But the idea that turnout this year will be greatly depressed because of the change of date is a bit of a reach, especially when comparing it to DFL turnout the last few cycles in which there wasn't really a contested primary. I don't want to go into too much detail because I have a post on exactly this subject I'm working on for later in the week, but if any of the campaigns are counting on depressed turnout to carry the day they will be in trouble.

The fact that Mark Dayton hasn't had less than a 10 point lead in any poll combined with his popularity among seniors, the most dependable group in mid-term elections, makes it a very difficult road for either of the other two candidates to pull out a victory.

hPVI in visual form

This is my attempt at putting all the hPVI information into an easy to understand graphical format, preferably one with pretty colors.

First I'll take you through the individual elements of the chart; the legislative seat chart, the hPVI chart and the house district spread chart, then combine those into one easy to understand Minnesota Legislative hPVI Chart With Pretty Colors.

What initially got the ball rolling in this direction was Joe's request for some sort of graphical chart of legislative seats.

…I wonder if there's a graphical way to show who controls the house seats in each of those districts.

This was the result of that:



Each cell represents a legislative seat, the numbers are Senate seats, and the corresponding letters are House seats. Blue means that seat is held by a DFLer, and red means it's held by a GOPer. The headings at the top are shaded blue because both chambers are controlled by the DFL; if one were controlled by the GOP the cell would be red.

Joe liked that chart but I wasn't quite satisfied, I wanted a way to integrate that chart with hPVI, which lead to the next chart, a way of formatting the hPVI information in a graphical way:



In this chart the headings at the top refer to the seat type, SD for senate district, A and B for that house district with the numbers along the side referring to the district number. The numbers in the cells are the hPVI numbers and the cells are color coded so that neutral districts are white, and the more Republican a district is the deeper shade of red it is, the more Democratic the more blue.

To that we can add a column for the district spread, the difference between the hPVI's of the two house districts, color coded from white to green.



Now we have two sort of similar looking charts that can be combined into one:



In one graphic we can see what party currently holds the district, what its partisan tilt is and how big of a difference in hPVI there is between the two house districts.

hPVI; House districts, sorted

Below is the Minnesota House district hPVI list.

As with the Senate district list the most democratic districts are significantly more partisan than the most republican districts. This is not surprising; as the house districts are further segmentations of the Senate districts we would expect them to retain some of the same partisan tendencies. You can see this in the graph, which looks very similar to the Senate hPVI graph.



And I've said it before, but I'll say it again, SD61 is really, really liberal. Really. For reals.

Here's the house hPVI list:

hPVI; Senate districts, sorted

Below is the entire Minnesota Senate district hPVI list, but before we get to that let's first take a look at some information about the list.

As you would expect the most democratic districts are significantly more partisan than the most republican districts. In fact a full 16% of the districts are Democratic districts that are more partisan than the most republican district. Conversely there are relatively few Democratic districts in the +10 - +20 range, while there are lots of republican districts in that range. You can see the trend more easily when the list is graphed.



The republican districts are more plentiful, but less partisan while the exact opposite is true of the democratic districts. The other interesting feature is just how many +10 republican districts the DFL currently holds. Even if you exclude Lisa Fobbe in her R+24 district there are six districts with an R+10 score or greater that are held by DFLers while the most Democratic district currently held by the GOP is SD41 with an R+5 score. This is to be expected in a Senate dominated by the DFL, but it's still somewhat surprising to see how far into GOP land the DFL incursion has penetrated.

And SD 61 is really, really liberal.

Here's the entire Senate hPVI list:

New Ras poll, GPI update

It's been awhile since the Minnesota Governors race was polled, SurveyUSA released the last poll on the race over a month ago, so we haven't been able to update the GPI recently. Before we get to that let's just take a quick look at the results of the Rasmussen poll.



All three DFL candidates polled better this time than last time Rasmussen polled this race, by margins of 3 to 5 points while Emmer and Horner polled slightly worse, by margins of 1 to 2 points. These are the same trend lines that SurveyUSA caught in their now month old poll, Emmer slipping and the three DFL candidates gaining. It's not hard to see how these trends will only be magnified after the DFL primary in August.

Now onto the GPI update. Just a reminder on the methodology:

The weighted average is an average of all the polls, weighted against a variable called "weight". This variable represents a combination of the Pollster Rating and the age of the poll. The newer the poll and better the pollster the higher it will be weighted. The age of the poll has more effect on the weight then the pollster rating, with the age making up 75% of the weight and the Pollster Rating making up 25%. However the age weighting will essentially fall off at 50 days and this will get adjusted as we get closer to the race in question.

There are only two polls that meet the 50 day shelf life criteria right now, the just released Rasmussen poll and the June SurveyUSA poll, so the GPI is simply a weighted average of the two polls. The only recent poll on the DFL primary is the SuveyUSA poll, so there will be no update to the primary GPI right now.

Here's the current GPI poll list:



And now the actual GPI:



All of the DFL candidates increased their wAve by about 3 points while Emmer held steady and Horner dropped a couple. The trend lines certainly look good for the DFL going forward.

Top 5 House districts by hPVI

Last week I posted the top 5 most partisan Senate districts in the state, according to hPVI, this week we'll look at the House version of those lists. First the top five most Democratic House districts:



Just like its Senate counterpart the Democratic list is all Twin Cities districts. In fact, the top 15 most Democratic house districts are located in the Twin Cities. And if you needed any more proof, SD 61 is really, really liberal. Really.

Here are the top 5 most Republican districts:



There's Tom Emmer, hailing from the fourth most Republican district in the state. The flip side of that is Margaret Anderson Kelliher, whose 60A is the seventh most Democratic district in the state at an hPVI of D+50. It's entirely possible that the two major party candidates for Governor will be from two of the most partisan house districts in the state, with a full 80 points of spread between their constituents.

I'll post the complete lists for both the house and senate soon.