Here’s what the Obama campaign’s issue models tell us about the Democratic Senate primary in Florida

Over the next few months, the Organizing Center will be diving deep into voter targeting data to breakdown some of the best political battles in 2016. We’re doing this with the help of L2’s Voter Mapping software that now includes Haystaq DNA data, also known as the Obama campaign issue models. These models allow us to score voters on a variety of issues, and they provide incredible insight into how an election might turn out.

One of the most compelling primary elections shaping up is in Florida. Republicans have struggled to recruit and have a divided field. Meanwhile, Democrats have fielded two heavyweights: outspoken progressive Orlando Congressman Alan Grayson, and young, measured, moderate West Palm Beach Congressman Patrick Murphy. The Grayson-Murphy matchup is shaping up as a great turf war between ideology, institutional support and geography. Whoever wins will be the favorite to take over Marco Rubio’s Senate seat in America’s biggest swing state.

We put the data to work to see what can be learned…


The first key question is, “who are the likely voters?” To get a sense of expected turnout, we looked at the last two major Florida Democratic primary fights. Kendrick Meek, Jeff Greene and a couple of minor candidates squared off in a 2010 Senate primary. Turnout was 910,393, and Meek was victorious with 57.4% of the vote. The 2014 Gubernatorial contest also featured two major Democrats vying for the nod: Charlie Crist and Nan Rich. Turnout was lower at 837.796, and Crist won in a landslide with 74% of the vote.

A total of 1,111,506 voters (still registered as Democrats today) cast a ballot in either one of these past primary races. We’ll use this cohort as our “likely voter universe.” It’s a higher turnout number than either the ’10 or ’14 race, but dynamics are different this time. 2016’s Presidential campaign cycle should create higher turnout, and both candidates are expected to be extremely well-funded. We’ll look to a target of one million votes — about 90% of the likely voter universe. For this reason, we’ll analyze all the likely voters, rather than trying to trim them upfront.

Next, let’s take a look at the basic racial demographics — key to any Florida Democratic primary. The following chart shows the ethnic breakdown, thanks to L2’s enhanced voter file. We can establish that about 700,000 of the 1.1 million likely voters are Caucasian-non-Latino, Asian, or unknown. We’ll call that the “white-plus” group of voters for the sake of paying special attention to the African American and Latino demographics. Both candidates are white, so this white+ group will be the starting point for understanding their bases of support.


Each candidate starts with a group of voters who are geographically and demographically closer to them. By identifying the strength of these natural coalitions, we can understand each candidate’s “base.”

The single biggest starting block difference between Grayson and Murphy is their media markets. Each has run major, expensive races in the past few years. But their ads are restricted mostly to particular media markets. Florida’s voters live in eleven different television zones, and wherever an elected official has advertised most, he or she will find the deepest support.

Grayson starts with an advantage here. His native Orlando cable zone has 210,079 likely voters compared to Murphy’s West Palm zone that includes 115,129 voters. That’s 18.9% for Grayson to 10.4% for Murphy.

West Palm Beach Media Market Democratic Voter Density

Murphy, however, can extend his base deeper. At just 33 years old, his youth is an obvious calling card that worked well for Rubio in winning this seat last time. Outside of the Orlando and West Palm cable zones, we can find 34,191 white+ likely voters who are of the ages of 18 to 40. They’re Murphy’s generation and will naturally speak the same language.

Murphy also has deep ties to the Miami — Ft. Lauderdale media market. His career almost began with a Congressional race in this market, and his family is well-known for South Beach real estate development. We’ve already given Murphy the young white voters, so we’ll give him the older white moderates in the Miami-Ftl media market.

In order to determine moderates versus liberals, we can use Haystaq DNA’s fiscal liberal score. This is especially useful because fiscal moderation compared to a full-scale liberal fiscal policy is the main flash point differentiating the candidates’ records. The Haystaq DNA scores every voter from 0–100, with 100 representing the most fiscally liberal voters possible. Traditionally, political targeting professionals will divide these scores into quintiles to group together the most conservative (0–19), the most liberal (80–99), and the three groups in between (20–39, 40–59, 60–79).

Looping back to 40+ year old white moderates in the Miami-Ftl media market… the Haystaq scores in the 3rd and 4th quintiles give us our Democratic fiscal moderates. (Almost no Democrats score in the lowest quintiles.) That leaves us with another 78,598 likely voters in the Murphy base.

We’re not quite done building Grayson’s base, either. If we’re granting some fiscal moderates to Murphy, it is only fair to grant the most progressive fiscal liberals to Grayson. Grayson is known across the state as a progressive, so we won’t restrict his media market. Another 14,201 white, 40+ likely voters live in the 5th quintile of fiscal liberal scores and outside of the media markets we’ve already discussed.

In sum, here’s a sense of the base voter universes for each candidate. Of course, they won’t each get 100% of their base votes. Instead, we’ll grant them a 75–25 split for base votes. Murphy’s base edges out Grayson’s, 227K — 225K.


This leaves us with a remaining white+ likely voter universe of 354,007. Breaking down this white+ “swing vote” cohort will provide our next level of insight into the race.

The biggest single demographic group of white+ swing voters is the elderly. 150,683 of the 354,007 are 70 years of age or older. Since we’ve already banked votes in Miami-Ftl, Orlando, and West Palm, we see a majority of these elderly living along the southern gulf coast. The Tampa media market holds 80,550 (54.5%). The Ft. Meyers media market, to the south of Tampa, is good for another 17,496 (11.6%).

Clustered in battleground turf and a demographic that both campaigns are sure to target heavily, we can’t just put these seniors into one box or another. We need to look deeper, and for these voters we’ll look to a very relevant cultural divide: religion.

L2 data allows us to breakdown the 150,683 swing white seniors by both religion and level of religious practice. A candidate named Patrick Erin Murphy is sure to appeal to older Catholic voters better than Alan Grayson. We’ll give him a 70–30 split among the 18,204 voters who fit the bill. This leaves us 132,479 swing white seniors.

Murphy also has a strong appeal to non-Catholic Christians, having attended a Christian school. Grayson, on the other hand, very much represents the secular progressive movement in Congress. Using Haystaq DNA data, we can break the remaining 132,479 into, “frequent church attenders,” and, “unknown” groups. The former is 69,650 voters, which we’ll give to Murphy with a 70–30 split. The latter is 62,829 voters, which we’ll give to Grayson by a less generous 60–40 margin.

Now our totals look different, with Murphy opening up about a 25K vote lead:


The remaining white+ voter universe contains voters of ages 41–69, who aren’t expressly fiscally liberal, and live outside the West Palm, Miami-Ftl, and Orlando media zones. This is a big, important battleground group of 195,965 votes.

These voters are the least prone to identity politics, simply because most don’t fall neatly into any specific group. Therefore, they’re the most likely to look closely at the issues.

One of the major disagreements between Grayson and Murphy is on foreign policy. Grayson has carved out strong positions against military expansion and foreign intervention. Murphy has been more of a mainstream Democrat, preferring a more “muscular” version of American international leadership. Murphy wisely appeals to the older swing vote in his home district as a Truman Democrat on security and foreign policy issues. But that might not help him in this race.

Here again, the L2 file with Haystaq DNA issue data can shed light. The file has a flag for positively identified, “military intervention opposers.” Of the 195,965 remaining white swing votes, 29,251 fit this bill, and we can bank 80% for Grayson. On the other hand, 11,931 are flagged as “humanitarian intervention supporters,” and we can safely give 70% of these votes to Murphy.

This leaves 154,783 votes up for grabs in the “white+ swing” universe, and narrows the gap for Grayson from ~25K to just about 11K.


Democratic Voter Density in Tampa

The final dividing line we’ll draw in the white vote is geographic. We mentioned earlier that Florida voters live in a whopping 11 different media markets. We’ve already divvied up the white+ voter in Orlando, Miami-Ftl, and West Palm. We have two major clusters remaining. First, the northern media markets: Dothan Alabama, Gainesville, Jacksonville, Mobile Alabama, Panama City, and Tallahassee. Second, the gulf coast markets: Tampa, and Fort Meyers shown below.

Our remaining age demographic is 41–69 year olds, and here’s the breakdown of votes up for grabs:

To recap the slicing and dicing of Florida’s white+ Democratic primary vote, here are the major groupings we’ve assigned. Note, each successive group contains no voters from previous groups. So when we label seniors, having already assigned fiscal liberals, the senior slice of the pie will not contain any fiscal liberals. Groupings follow clockwise:


Now we can take a look at the remaining Latino electorate of 67,920 likely voters. Normally, we’d want to avoid volatile polling data in favor of known issue support data and understanding fundamental dynamics. This is especially true right now, because the race is so early and there has been almost no advertising.

However, September’s Public Policy Polling (PPP) report on the Murphy-Grayson matchup showed a huge Latino vote lead for Grayson: 43–13 with 44% undecided. This is possibly because Grayson’s Orlando media market contains four times as many Latino voters as does Murphy’s West Palm market.

Another interesting trend leads us to believe there’s a there, there. With other demographics, Grayson has a “net unfavorable,” rating meaning more voters think negatively of him than positively. However, he has a big lead with Latinos, especially compared to Murphy (who is nearly even). Even if Murphy earned an even split the remaining 44% of the Latino vote, we can still bank 65% (non-Orlando/West Palm media markets) for Grayson. This puts Grayson ahead by almost 9K votes.


The final demographic left is 236,224 African American voters outside of Orlando and West Palm media markets. Neither Murphy nor Grayson has a natural base with the black vote, though Grayson has a sizable African American community in his geographic base. This means we’ll have to dive a bit deeper into the composition of the community.

First, we can determine from Haystaq DNA flags that these voters are very liberal. 98% fall under the “strong democrat” flag. 94% support minimum wage increases and 91% support unions. 87% have the “climate change believer” flag. With Grayson as the more liberal candidate, we might anticipate an advantage. On the flip side, we might also anticipate Murphy being smart enough to take a very progressive tone with black voters. Result? We’ll need to understand more than just issue scores.

The second characterist is poverty. Of our 236K remaining black likely voters, L2 data tells us over 40% are in the bottom income quintile of Florida residents (and Florida does not have a relatively high median income). We can also see that 100,887 (42.7%) are in the Miami-Ftl media market. But a third stat from L2’s voter mapping might give real insight: this group is overwhelmingly female — 151,318 (64%) are women.

To sum, the typical Florida black primary voter is female, near Miami, lower income, and very liberal. Turning back to PPP’s poll, Murphy may have an edge. Geographically he’s closer than Grayson, but he also has much high favorables among women voters. Knowing this, we’re willing to give a conservative 55–45 split in the black female vote to Murphy.

Murphy shoots back into the lead, with an anticipated 6K vote lead:

What left to discuss? Just the remaining male black vote of 84,798, which may very well determine the election. Again, geographic targeting may well make the difference. The battle for these votes will likely be won through a field campaign, pounding doors, working the phones, and hitting the pews.

Let’s take a look at where these voters live:

Our fearsome Democratic candidates will have to pay attention to each of these zones to win the votes remaining. The difference could come down to a Jacksonville field effort or smart media buy Tallahassee.

Here’s the total remaining “up for grabs” universe:


The key to the election is now simple: Grayson will have to make up about 7,000 votes in these three, key, nearly equally-sized voter cohorts to defeat Murphy.