Category Archives: Predictive Analytics

Analyzing “First In The South” Voter Turnout Programs


This year’s “First in the South” presidential primary in early-voting South Carolina saw a massive 23 percent increase in turnout from 2012.  It also saw 215,082 first-time voters – comprising nearly 30 percent of the GOP electorate.

Prior to the election, various efforts were undertaken to identify potential first-time voters who would support presumptive GOP nominee Donald Trump – and to target these potential voters with multiple communications aimed at increasing their turnout.  Market Research Foundation was asked to verify the efficacy of this first-time, pro-Trump model – and provide back-end analysis of its impact.

Did the model work?  Absolutely.

Of the 122,000 modeled voters who received seven communications from the pro-Trump Super PAC during the final twelve days of the race – 25,063 voted for an impressive 20.5 percent turnout rate.  Among the control group of 20,000 voters, only 1,760 (or 8.8 percent) cast ballots.  The 188,259 remaining first time-voters represented 8.3 percent of the 2.2 million eligible voters in the state (minus repeat GOP voters).

Here’s the breakdown:

Continue reading Analyzing “First In The South” Voter Turnout Programs

Immigration Exit Polling


Electoral science is all about digging deep – uncovering elusive wisdom from within reams of seemingly divergent data.  In fact those who succeed in this business do so precisely because they know how to dig – and more importantly, where to dig.

Sometimes, though, we don’t have to get our hands dirty to find the answer.

Sometimes the answer is sitting right there on the surface – staring us in the face.

Take the big victories registered by Republican presidential frontrunner Donald Trump in New Hampshire and South Carolina – which held the first two GOP primary elections of the 2016 cycle this month.  Granite State voters have traditionally been viewed as more centrist – i.e. more friendly to GOP candidates residing on the left end of the party’s ideological spectrum. Meanwhile voters in the Palmetto State are viewed as far more conservative – especially on social issues – and far more eager to support candidates who appeal to their evangelical beliefs.

Neither of those electoral universes would seem to be especially receptive to a candidate like Trump – yet he won convincing victories in both states.

In New Hampshire, Trump won 35.3 percent of the vote – easily outdistancing centrist Oho governor John Kasich.  In South Carolina, Trump captured 32.5 percent of the vote – giving him a ten-point edge over both Ted Cruz and Marco Rubio, a pair of candidates who made overt appeals to socially conservative voters (and attacked Trump for his social liberalism).

What gives?  Immigration, that’s what.

According to exit polling in New Hampshire, 65 percent of GOP primary voters supported Trump’s proposal to temporarily suspend Muslim immigration into the United States.  In South Carolina, exit polling showed 73 percent supported Trump’s plan.  Those are compelling majorities.

Continue reading Immigration Exit Polling

Big Data Is Back!


Labor and Tory leaders in Great Britain have retained the services of senior Barack Obama campaign strategists in an effort to duplicate the data-driven success of his two victorious presidential campaigns.

According to analyst David Richards – writing for Tech City News – party leaders have come to realize “the ability to collect and analyse publically available data on a large scale allowed the Obama team to model behaviors before coordinating and targeting communications accordingly.”

This gave Obama’s campaign – which had bigger electronic and social media databases – invaluable insight into which messages were more likely to motivate people to go to the polls and cast their ballots.

“Not only did they have voters’ email address, they also had phone numbers, where they were registered to vote, a decent estimate of their household income and whether they’d opened a credit card recently,” Richards report noted.  “Obama’s camp knew how many children voters were likely to have and what they did for a living. And he knew what time of day people tended to get around to plowing through emails and respond to messages.”

Polls in Great Britain show an exceedingly tight race – with the distinct possibility of a hung parliament.  A YouGov poll found the Labor and Tory parties tied at 34 percent support, while a TNS poll had the Tories ahead by just one percentage point (well within the margin of error).

“Analytics can also make each pound work harder, with parties better able to focus on targeting the voters they need, helping limit overspending,” Richards added.  “Imagine that 10 seats will decide the election – by knowing where seats these are and what issues will sway the electorate, parties can save a fortune in terms of pound-per-vote spending.”

Will “big data” deliver the day for one party or the other?  Or will the efforts of the Obama election scientists cancel each other out?

Tory leaders enjoyed a decisive financial advantage down the homestretch, out-raising Labor by a 10-to-1 margin over the final week of the election.

Will they target those resources effectively – using “big data?”

We’ll find out soon …


Race And Age In America

Like it or not, a person’s age and race play huge roles in determining how they might vote on a specific candidate or certain issue.  And while we’ll leave it to sociologists enlighten us as to why that is the case – it’s important not to discount how these and other determining factors impact outcomes at the ballot box.  It’s also important to assess what methods or messages can be employed to maximize – or minimize – these impacts.

Breaking down demographics is central to our work here at Market Research Foundation (MRF).  Just click through a few of the articles on this website and you’ll see multiple references to the “percentage of black voters” or the “number of young voters” or the “habits of white voters” or the “trends of older voters.”  And ask any organization modeling likely voters for an upcoming poll or grassroots election effort and they’ll tell you: Age and race matter.

The first step in that process?  Determining a baseline for these specific demographics – which is where a new interactive feature from comes into play.  Using 2013 data from the U.S. Census Bureau, the interactive map highlights the racial composition – and age breakdown – of every single county in the United States, providing regional, state and national demographers with an incredibly useful visual tool.

For example it shows where black voters are most concentrated …

(Click to enlarge)

black voters

(Map: Via)

And where Hispanic voters are most concentrated …

(Click to enlarge)

hispanic voters

(Map: Via)

The map’s interactive feature also enables users to mix and match categories – letting them compare different age ranges with different racial breakdowns in each county.

Pretty nifty, huh?

Obviously interactive maps like this can always be improved upon.  For example, data from prior Census surveys could be uploaded in an effort to show shifts in age and race.  Meanwhile county-wide historical electoral data – such as is found on Dave Leip’s U.S. Election Atlas could be incorporated.  And of course regional and local candidates and campaigns would need to see all of this information broken down by precinct, not just by county.

Still, this map represents a nice baseline from which to begin an examination of how ages and races are distributed across America.


Early Polling: Don’t Believe The Hype


Recent news flashes from the National Republican Senatorial Committee (NRSC) touted a series of Quinnipiac University polls taken in Florida, Ohio and Pennsylvania.   Part of a broader 2016-themed “swing state” survey, the polls found freshmen GOP Senators in these three states – Marco Rubio, Rob Portman and Patrick Toomey – enjoying favorable reelection prospects.

According to the data, Florida voters approve of the job Rubio is doing by a 47-35 percent margin – and believe he should be reelected to a second term by a 44-37 percent margin.  In Ohio, voters approve of the job Portman is doing by a 40-21 percent margin – with the first-term Senator enjoyed a 37–28 percent “deserves to be reelected” spread.  Finally, Toomey’s approval advantage in Pennsylvania stood at 43-25 percent, with his reelection margin at 37-29 percent.

This is good news for the GOP, right? After all Barack Obama won these three states in both 2008 and 2012 – so the fact three freshmen GOP Senators are poised to hang onto their seats is a positive thing for the party, correct?

Yes … and no.  As one pollster noted, Toomey’s numbers in particular “aren’t very different from then-Sen. Mark Udall’s numbers” in a June 2013 Quinnipiac poll.  And we all know what happened to Udall.

“It’s wise not to read too much into early surveys, remembering that national atmospherics and state-specific factors will have a substantial impact on how races develop,” pollster Stuart Rothenberg wrote.  “Don’t ignore where candidates start. But always remember that where they end can be very different from where they begin.”

This isn’t to say early polling is useless.  Surveys such as this one provide important mile markers for pundits and press (as well as prospective challengers).  Also, for candidates and campaigns internal early polling represents a chance to test messages that could help them define upcoming races on their terms.

But “national atmospherics” and state-by-state influencers are both fickle forces … capable of shifting on a dime and developing velocities all their own.  Just ask Mark Udall and countless other elected officials who have seen early leads evaporate – turning likely victories into surprise defeats.


Oregon’s Independent Streak


Last week this website unveiled a new report looking beyond partisan identification – or the percentage of the electorate identifying itself as Republican, Democratic or (increasingly) independent – and delving into the ideological matrices that drive people’s voting behavior.

To read more about the study, CLICK HERE.  To read the study itself, CLICK HERE (.pdf).

Prior to this nationwide study, Market Research Foundation (MRF) conducted research on party loyalty in South Carolina.  The Palmetto State is reliably Republican – voting for the GOP presidential candidate in the last nine presidential elections.  But South Carolina is home to a strong (and emerging) independent streak – one you can read more about in our report HERE.

All the way across the country is Oregon – a reliably Democratic state that hasn’t voted Republican since 1984 and which elected Barack Obama by 16-point and 12-point margins in 2008 and 2012, respectively.  According to The Oregonian, voter registration in the Beaver State in advance of the 2014 midterm elections has surged to 2.2 million people – a record for a midterm election (and just 7,000 voters shy of the 2012 record).

Naturally record midterm voter registration means an increase in both major party bases, correct?

Wrong.  As Oregon’s electorate is expanding, the share of Democratic and Republican voters is declining.

“Voters who are non-affiliated or registered with one of the minor parties — particularly the Independent Party — now constitute 32.1 percent of the electorate, a record for Oregon,” the website reported this month. “For the first time in at least modern Oregon history, Republicans have now dipped below 30 percent of the electorate while Democrats are at 37.9 percent of the electorate.”

By contrast, forty years ago the independent percentage of the electorate stood at only 4.2 percent.

In red states and blue states – and the “swing states” in between –  there is a clear trend toward voter independence.  Smart campaigns would be wise not just to acknowledge it, but understand the ideological underpinnings of this movement.

Early Voting In Iowa


Long before all the ballots are counted on November 4, the nation’s political focus will have already begun shifting to the 2016 election cycle – with particular attention paid to presidential primary contests in early-voting Iowa, New Hampshire and South Carolina.  In both parties candidates will declare their intentions, frontrunners will be anointed and political fortunes will begin to rise and fall.

This long and winding road will eventually lead us to the 2016 national conventions – and the selection of a pair of major party nominees for the presidency.

Given their poll position on the primary calendar, voters in Iowa, New Hampshire and South Carolina will receive disproportionate face time with dozens of presidential aspirants. But they will also receive disproportionate analysis from political operatives and electoral scientists looking to ascertain what motivated them to choose certain candidates over others – and increasingly, “when.”

Early voting is all the rage in contemporary campaigning – and nowhere is early voting (and the micro-targeting electoral science accompanying this distinctly modern political phenomenon) more important than in early voting states.

“It used to be you just kind of let the electorate do what they were going to do if you were a campaign. We’d find out Election Day,” Iowa’s Ann Selzer recently said. “Well, now they’ve found out If we can get all of these early ballots out there, if we can get people to a poll early, we can recruit those people who don’t really care, but they’re going to vote for our candidate — and those votes are now there.”

Selzer is the president of the company responsible for conducting The Des Moines Register‘s “Iowa Poll.”  She and several other Iowa political experts gathered recently for a discussion of the 2014 election cycle – in which early voting was a hot topic.

That’s understandable.  Early voting in Iowa began on September 25 for the November 4 election – and given that the state is home to a competitive gubernatorial election and U.S. Senate race, both parties are making a major play for these votes.  Democrats were tremendously successful at their early voting efforts in 2012, and this year Republicans are looking to match – or exceed – those gains.

“Campaigns are certainly getting more sophisticated in their use of the state voter file, in terms of finding likely voters,” University of Northern Iowa professor Chris Larimer said. “They’re also making campaigns a bit more personal because they know that a more personal touch – that can be done through direct mail, but it has to be done carefully – can also affect a campaign. So, I think they’re changing strategies and we know that early voting is going up all across the board.”

Indeed.  According to CNN, the GOP is already reaping the reward of a $1 million investment in early voting infrastructure in Iowa.

“In just the last week, Republicans have requested absentee ballots at a faster rate than Democrats — more than doubling their count compared to just a 40 percent increase for Democrats in the last 10 days, according to numbers provided by the Iowa Secretary of State’s office,” the network reported. “Registered Republicans have also mailed in their early ballots at a faster pace than Democrats. And while both parties have gotten their voters to submit more ballots than in 2010 so far, early Republican votes have more than doubled from 2010 while Democrats have only posted a 36 percent bump.”

In fact Democrats have dispatched First Lady Michelle Obama to Iowa in the hopes of providing an early voting boost for their candidates.

“Campaigns love early voting,” Drake University professor Art Sanders said, although he argues its real benefit is offering operatives (and political scientists) a chance to see which methods are most effective in moving unlikely voters.

“It allows them to tailor strategies in much more sophisticated ways,” Sanders said. “It’s very effective from their perspective.”

For example, one method that’s reportedly been quite effective in Iowa is to promise early voters that they won’t receive any additional political mailings once their ballot has been cast!  Talk about a powerful motivator to those who are fed up with the incessant pre-election bombardment of political literature.

Of course while Iowa’s 2014 elections are critical in terms of deciding which party will control the U.S. Senate, Democrats and Republicans are investing heavily there because they know research on what motivates early voters in Iowa will be indispensable come 2016 – in the presidential primary elections and the general election in which Iowa is likely to be a swing state.

Lessons learned in this race will pay huge dividends down the road.