How ‘Big Data’ Gives Pro-Life Candidates a Big Advantage (and Helped Trump Win)

By | 2017-06-02T18:30:05+00:00 January 28, 2017|
Print Friendly, PDF & Email

 

The biggest problem in Republican politics is that the power brokers have a disincentive to accountability. If you don’t know what works or what doesn’t, you can’t be held accountable for failure and you can always claim that someone else is why you lost. Ignorance is bliss.”

With those words from a prominent political consultant in December 2013, I was introduced to one of the biggest problems in conservative election efforts. With a decision to follow the data wherever it led, I pledged then to make sure the national pro-life movement began a course change with the goal of influencing every Republican effort in the years ahead.

Though this effort is still a work in progress, giant strides were made in both 2014 and 2016 which led to significant efforts in at least nine battleground states on behalf of Donald Trump. These data driven efforts are the gold standard of efficiency and effectiveness and the entire Republican world will be improved as these methods spread.

I had given up on political campaign work in 1992. Too many dishonest and self-dealing campaign professionals drove me away. So it was easy for me to accept this idea that the election professionals preferred ignorance and their own instincts to science.

From the late 1980s until 1992, I had been closely involved in political campaigns. In 1992, I worked for Pat Buchanan to challenge President George H. W. Bush. I admired Pat and stood fully in agreement with his positions.  I remember watching the 1992 Super Bowl with Pat and Shelley and about ten of Pat’s NH team in a house in Manchester New Hampshire.  As I looked around me at the other paid staffers, I saw man after man who was quite explicit in saying he didn’t actually believe in Pat’s ideas but was just using Pat as a stepping stone. It was clear to me that these opportunistic men would be the power consultants of the future. I turned away later that year from a career plan in campaign management and focused instead on new efforts in the work related to my Catholic faith and conservative education efforts.

Searching for a Secret Weapon

I found myself in late 2013 recruited by friends to lead an innovative pro-life plan to spend over $1 million in the race for governor in Texas exposing the radicalism of the Democratic candidate. We weren’t sure whether we would need $1 million or $2 million or $5 million but we were trying to make sure that the truth about liberal Democrat Wendy Davis was known to pro-family and pro-life Democrats.

Knowing how long I’d been away from campaigns, I began to reach out to any and every expert to find how to make our work most effective. I figured I’d find other efforts to copy. I was looking for a secret weapon in strategy.

We quickly learned that there was almost no history of this level of pro-life spending. Not only was the level of spending unprecedented, the bigger truth was there was almost no reliable research into what worked and what did not. Only two organizations, National Right to Life and the Susan B. Anthony List, had ever spent amounts this big, but before 2012 those efforts were rare and spread over many races. Like the rest of the Republican world, election plans were made entirely by trusting experienced campaign experts who just followed their “gut.”

A Lot of Exaggeration and Lies

As my team studied Republican election efforts from past years and studied other conservative efforts across the country in that 2014 election, I quickly recognized those old self-interested campaign workers now running the show. And all too often I heard unbelievable claims by them. Each one claimed to be on the cutting edge of political technology. Each one was using “microtargeting” or “data mining” or whatever other sexy description made them sound most efficient and effective. Since I was no expert, I could not easily judge these claims but again and again I found bold claims with little evidence to support them.

It would take a few months and a remarkable academic researcher before I would learn how much exaggeration and dishonesty was out there. Sometimes the best position to be in when you are searching for wisdom is total ignorance because you have no prejudices. From that position I began to look for a pollster or any other researcher who could tell me how to design the most effective advertising. Not knowing anything I asked lots of questions of many of the nation’s biggest polling firms and again and again found that they did not have answers for me.

The problem wasn’t that they knew nothing, but that they didn’t know the one most important thing. One pollster told me that his polling showed that Hispanic voters were overwhelmingly pro-life and that they did not vote based on that fact. This were not news to me. What I wanted to know was “what would change that.” It was this question that was always a problem.

Campaigns would often we able to say which people liked their candidate and which ones did not, but they had no data on what messages were shown to be most likely effectively to change that. It sounds simple, but it is of course a very difficult thing to determine.

You cannot just say to a man, “I know you think X, but what could I tell you that would make you think Y?”

One Unusual Expert in Political Psychology

The future of Republican political technology became clear to me in December 2013 when I read an article by Dr. Adam Schaeffer in an elections trade journal called Campaigns & Elections. Schaeffer and his team run a political research shop called Evolving Strategies. Schaeffer explained how he had been able to use a new kind of testing called “randomized, controlled experiments” to answer the question of what would change the mind of a voter. His testing could predetermine what kind of advertising was most likely to persuade opposing voters and this could help to win a race. But his article showed how the campaign staff of the Republican candidate would not accept the data. They favored their own instincts and experience. I called Schaeffer and quickly connected the dots. Here was the secret weapon.

This conversation with Schaeffer, I believe, marked the beginning of what would be a Republican election technology revolution led by the pro-life movement.

Liberals Have Dominated Good Election Data Usage

The question I had about how to effectively change minds revealed something that, in fact, was not “new” at all. This entirely  effective solution had been discovered more than a decade before in the world of academic research. Indeed, this explains the dominance of the political left in campaign strategy during this time.  They have been using this science to win elections for over a decade because their partisans were behind these discoveries.

Starting in the early 2000s, political scientists Alan Gerber and Donald Green of Yale University began to use these randomized controlled experiments to study the relative value of one message versus another. The experiments are designed to be random so that they effectively show a trend that can be extrapolated to large populations. They are controlled in ways that are designed to measure real changes of opinion without the study subject being aware what the goals of the study are.  The key to these kind of tests is that they follow established statistical rules making their results trustworthy when applied to the greater population.  Amazingly, this is generally not the case in existing Republican research efforts.

Much like the ways that a drug study avoids the placebo effect and tries to measure only real provable results, these experiments try to measure what message is most persuasive in real life settings.

And they measure not only what message works best, but also help determine which voters are most influenced by such good messages. The combination of 1) who is persuadable and 2) what persuades them best is simple and can change campaign results.

Quickly, the liberal friends of these researchers, Gerber and Green, caught on to these valuable techniques and how they eliminate the problem of bias in polling and most other testing. With money from the massive union election arm and in response to the loss of Senator John Kerry, liberal organizations brought these scientific researchers into their election planning beginning in 2005.

An organization called the Analyst Institute was the guide and advocate of most of these left wing experiments and they have become famous for their successes in election plan development.

Major Democratic victories in 2006, 2008 and 2012 as well as many small victories at the state level in those years were due, in great part, to these scientific researchers and their adjustments to Democratic strategy.  Today, these methods, the modeling of the electorate and message pre-testing using them are ubiquitous in liberal election planning.  But the Republican world mostly ignored these developments. Powerful consultants felt they didn’t need advice from liberals and young academics.

An Accidental Discovery that Life is the Most Powerful Issue

In that 2013 Campaigns & Elections article, Schaeffer describes how he stumbled onto the power of the pro-life issue to persuade Democrats while doing one of these experiments in regard to the 2013 Virginia race for governor:

Evolving Strategies and the Middle Resolution PAC conducted experimental research that suggests an aggressive attack on McAuliffe for supporting ObamaCare was ineffective at best and counter-productive at worst. An attack on McAuliffe’s business record possibly helped, but was anemic. What moved the voters most was an attack on McAuliffe’s positions on abortion; a single phone message emphasizing McAuliffe’s support for unrestricted, late-term, and taxpayer-funded abortions shifted support a net 13 to 15 points away from McAuliffe and toward Cuccinelli. The cost per vote here was a remarkably cheap $0.50 per additional vote, and even less expensive still when targeting the most persuadable segment of the electorate. A topic declared radioactive by nearly everyone, locked away in secure storage behind a blazing Hazmat warning by the Cuccinelli campaign, appears to have been a powerful weapon for the Republican ticket that could have substantially closed the gap, and possibly even won Cuccinelli the election.

Schaeffer was doing this research for a PAC uninvolved in the pro-life movement and he himself had not anticipated that abortion ads would work as well as they did. It was an accidental discovery.

“Though I was surprised by the finding that marginal Democratic voters and especially women were powerfully persuaded to abandon the Democratic candidate based on his extreme position on abortion, it made sense in retrospect,” Schaeffer said.

Many of these voters are not political junkies. They are not all that interested in politics. And they are not highly committed to the Democratic party. Their support for Democrats is primarily habit. They come from old Democrat families and take for granted that the Democrats are the party of the little guy. But these are the regular folks who still go to church and are parents with children. More important, these voters have almost no awareness of the detailed abortion position of either side. They are surprised and even shocked to hear that the Democrat would allow abortion up to the date of birth. Their virgin ears are often repelled by this new bit of data when they hear it and it is so shocking as to make many of them consider abandoning ship.

This data came from a randomized controlled experiment which got past the assumptions of the “experts.” That is the value of this experimentation. One can begin with assumptions but you test them with real voters and often you get shocking surprises.

Schaeffer has conducted over 20 such experiments on aspects of pro-life election advertising and the power of abortion advertising has been confirmed again and again in many different geographic locations and many different election settings.

This body of research though not particularly large constitutes not only the largest body of research ever on abortion in election politics, but probably the largest body of data on any one election issue.

When we applied these methods to our planning in Texas in 2014, the results of our ads were simply shocking. Our testing led us to hone in on Spanish speaking voters. Our tests also told us to be very bold and direct, especially about how Hispanics suffer more in terms of percentages from abortion. The result was truly one of the most impressive election shifts in any recent election.

Our ads began in September 2014 through robocalls into targeted Democratic homes and through millions of internet banner ads and finally a massive blitz on Spanish language radio. The results were dramatic: after nine months of almost no change in Hispanic support from our daily tracking polls, we saw an 18 percent shift in the Hispanic vote in six weeks. We went from a September average of 27 percent support for the Republican candidate statewide to an election day result of 45 percent support. This amounted to a 4 percent overall increase in the Republican’s vote.

Even before this exciting result, D.C. pro-life groups took notice of Schaeffer’s reports and his ideas were utilized in an election effort in the Iowa Senate race in 2014 that elected U.S. Sen. Joni Ernst. Further experiments there with direct mail pieces that showed again how one might reach out to Democratic women and see large numbers persuaded over the abortion extremism of Democratic candidates.

The Susan B. Anthony (SBA) List had already profited greatly over the Obama years and had gone from extremely modest spending in congressional races of just a few thousand dollars per race in the Clinton and Bush presidencies to a multi million dollar campaign in 2012 and 2014. They had rocketed in size past the only large organization, National Right to Life, thanks to a smart focus on women candidates and an aggressive response to punish the pro-life Democrats who, in 2010, had betrayed the pro-life movement by supporting Obamacare. The 2014 cycle was a good Republican election year and SBA had successful efforts in several states. With this new information from Schaeffer’s research, they had evidence of ways to use that larger funding far more effectively.

I remember talking about the possibilities ahead in July 2014 with SBA’s president, Marjorie Dannenfelser. Both of us remarked how much could be done to change the landscape of elections in the future. Marjorie and I committed then to do all we could to advance the cause together for 2016.

Schaeffer and I also knew we had to do more testing. We knew we’d learned many useful things in 2014 but that we could have done much much more if we had only had more time and money. We knew the possibilities for the 2016 presidential election were massive. We picked three likely battleground states for 2016 with high Hispanic populations where we thought a Hispanic effort might help: Colorado, Nevada and New Mexico. We also identified Wisconsin, Michigan and Pennsylvania as three states with large numbers of white “blue-collar” voters who our testing had identified as very open to persuasion by our methods. We created a Super PAC called Valor America  in the Summer of 2016 for our pro-Trump advertising and have now created another organization called the National Life Persuasion Project for our ongoing data gathering efforts beyond 2018.

The SBA List chose to pursue large efforts in Florida, North Carolina, and Ohio.

What’s Happening in 2018

It will take some more years before pro-life election efforts cross $50 million or even $100 million as is spent in every cycle defending the Second Amendment and in support of lower taxes, but with only 10 percent as much money the 2016 efforts of Valor America and all pro-life ads for Trump had a giant advantage of being designed based on proven messaging and being targeted to proven persuadable voters.

The results in those nine states are known now and of course Donald Trump won. Many people and their organizations will take credit for what they did. It is not possible to determine what efforts were most important in any state, but across more than 50 national organizations on the Republican side that engaged in major support for Donald Trump, there are only two whose efforts were designed according to these randomized controlled experiments.

It will take time for the other campaign professionals to abandon their instincts in favor of such testing, but we are well on the way to that result. And this is because donors will vote with their checkbooks. People with a lot to give value success and proof.

As we look out toward the 2018 election, I am already crafting a much larger plan for my expanding team to make our National Life Persuasion Project not only the biggest research organization in the Republican election world but to expand our research into some of the other possible issues that seem to have worked for President Trump.

There are 23 Democratic U.S. senators up for reelection in 2018, along with two liberal independent senators. A quick look shows real potential in anywhere from 10 to 13 of those races.

Even in a state with a pro-choice majority, one can advertise directly to identified persuadable pro-life Democratic leaning voters and maybe see a close race tipped the other way.

In this new digital world there are more and more ways to get an ad in front of a specific voter while avoiding the backlash that would come if his next door neighbor saw the same ad. Combining that ability with cutting edge data about who thinks what is part of the future of winning campaigns.

In ten years one hopes that every Republican organization will be doing these things and, when they are it will be thanks in no small part to the efforts of a very small number of pro-life activists with the cause of righteousness on their side and the hard work of a once obscure scientist with a Ph.D. and the willingness to go where the data took him.

About the Author:

Joseph Arlinghaus
Joseph Arlinghaus is 46 years old and lives in Fort Collins, Colorado with his wife Elsy, the younger three of their six children and a 110 pound Great Dane named Juno. Joe and Elsy are 1991 graduates (in Politics and English respectively) of the University of Dallas where their three oldest children are attending simultaneously in a conspiracy to bankrupt the family. Joe is the President of a large new pro-life research organization called the National Life Persuasion Project and a national pro-life election effort called Valor America Super PAC. His partnership with a political research firm called Evolving Strategies has resulted in the largest body of research data in existence on how voters respond to a variety of abortion related election messages. His 2016 election efforts on behalf of Donald Trump were focused on outreach to pro-life Hispanic voters and pro-life white blue-collar Democrats. Joe served as a consultant in 2015 to the Washington based Susan B. Anthony List. In December 2015 he was the first major pro-life leader to endorse Donald Trump for President and was one of the most vocal advocates of Donald Trump's credentials on the pro-life issue throughout the early caucuses and primaries.
  • ricocat1

    A long article which contains some good ideas which the GOP can use to create a Senate super-majority in 2018. One galling quibble: a writer looks lazy when they use the word “hone” when the proper word should be “home”.

  • Joseph Arlinghaus

    I assume we can fix the typo. I’m not that great a writer. I might have written hone without thinking. My late grandfather would scold me for my poor vocabulary. He continually lectured news anchors through his TV screen about their misuse of words like “momentarily”. When they said “we’ll be back momentarily” he would say to the TV, “we can only hope it’s true.”
    I hope the issue of proof and its power to win is clear. It is an amazingly simple thing that is shockingly uncommon. What is stranger is the great lengths people go to in their election work to do things that look like proof but which are nearly always guilty of following statistically invalid methodology. For the sake of an example allow me to tell you about a pollster’s efforts to help me with such a study.
    Before I learned the techniques I now use, I ran some typical negative ads about a year out as an outside PAC for a major candidate in a statewide race in Texas. The ad purchase was relatively small ($20,000 of radio). A week before the ad, the official pollster for the Republican candidate ran a statewide poll of voter opinions regarding that candidate. Two weeks after, they repeated this poll. Their conclusion was that our ads were fully responsible for an almost 25% shift in the favorable unfavorable ratio of the Democratic candidate in the Rio Grande Valley area where we had run the ads. This is an area of far south Texas centered on the city of McAllen
    Naturally I loved this conclusion and initially I believed it.
    Much later, my data partner Dr. Schaeffer tore apart this conclusion:
    1. Each poll had polled 1000 people statewide which implies that there would be maybe 100 people polled in Dallas and another 100 in Houston but that a small town like McAllen would be lucky to have contained more than five of those people polled. Clearly a poll of five people has giant problems.
    2. He said this sort of pre-poll/post-poll comparison is rarely useful for analysis of advertising because there are thousands of other things happening in the world simultaneously and it is simply impossible to determine tht a certain ad caused the shift. This is the problem of “correlation does not prove causation.”
    At some point I bumped into one of the owners of this polling firm who readily admitted that the poll conclusions were “crap” and he knew it.
    Yet this is going on all the time.
    It is extraordinarily hard to do any sort of testing, especially in a noisy race where hundreds of efforts are happening simultaneously, that can prove after the fact that a project caused a win.
    But what we can do is pre-testing where simultaneously we show large groups of voters multiple possible ads and determine which group responded most favorably to the ad. This testing is what I am advocating and by doing it, we could weed out easily 75% of all crappy ads on our side. Those ads hurt our more successful efforts and waste precious resources at the same time. More bang for less bucks!

    • Northbrae

      What a bonus to have the author reply to his own article. Wonderful stuff, Mr. A. The fewer the bucks, the better. (Wink wink; thumbs up.)

      • Northbrae

        Also galling — I swear I am not lashing out; I just hate it — is the linkage of ‘everyone’ with ‘their’. I am not attacking Rico. It is just that I am galled.

        • Joseph Arlinghaus

          The gender neutral world is infecting me. Or again maybe just my lack of good schooling. I agree with you. Should be “his”

      • Joseph Arlinghaus

        I more than anyone want to raise giant sums to do giant work, but money that is wasted on inefficiencies is offensive to me. One of the best examples in the consultant world is the obsession with large and expensive TV campaigns. No advertising is as ineffective generally and it is the most costly. The ineffectiveness comes from the fact that it is not at all targeted. It is broad cast. Everyone sees it, whether he be friend or foe (like my agreement in number there?). But the typical consultant is both paid more for a bigger budghet and knows that the TV impresses the candidate and the donors. So they blow giant sums on this big bucks little bang method.

    • AEJ

      It’s okay Joseph. ricocat1 also erred in his post to you:
      “A long article which contains some good ideas which the GOP can use to
      create a Senate super-majority in 2018. One galling quibble: a writer
      looks lazy when they
      use the word “hone” when the proper word should be
      “home”. The sentence “Our testing led us to HONE in on Spanish speaking
      voters” is very jarring. Hone means to sharpen, home means to zero in
      on”
      To err is human.