Polling v. Modeling: Can A Ballot Initiative Win In 2016?

Radically different polling and modeling numbers were reported at an LGBT community forum Dec. 7 on the topic of a ballot initiative for LGBT and women's civil rights. Polling data from Fair Michigan, the ballot question committee led by civil rights attorney Dana Nessel, shows strong support across all demographic sectors. But modeling analysis reported by Amy Mello from Freedom For All Americans paints a much bleaker picture. She reported data that shows a ballot initiative would lose next year in Michigan.
This discrepancy prompted a call from Mark Schauer, former congressman and 2014 Michigan gubernatorial candidate, for deeper analysis of all the research and empirical data available. He said that campaign organizers have to know and agree on what is the real starting point so the community can rally behind a unified plan to win in 2016. The meeting occurred on Tuesday in Ann Arbor.

What's In the Numbers?

Polling asks potential voters directly how they intend to vote, either over the phone, in focus groups or in one-on-one conversations. Nessel reported the results of a poll conducted by The Glengariff Group, a highly respected market research firm for political issues, that showed the proposed ballot language reached support of 68 percent YES, 24 percent NO and 8 percent Undecided. They found strong support across a broad swath of the electorate including Democrats, Republicans, Independents, all age groups and church-goers. Voters said their top reason to amend the Michigan Constitution was to prohibit discrimination, and 74 percent said they believe discrimination exists based on both gender identity and sexual orientation.
Modeling is a sophisticated analytical tool used by political campaigns to predict how a specific electorate will vote in a given election. It uses many of the same data mining techniques that marketing companies use to determine who will buy a company's products. FFAA did a modeling analysis in mid-November that found likely voters in Michigan in 2016 would vote YES only 42 percent of the time.
Amy Mello, public engagement director at FFAA, flew in from Denver Dec. 7 to present the group's findings. In a phone interview with BTL the next day, Mello explained the methodology and findings from FFAA's modeling.
"First, it is important to know that Dana's numbers are not wrong," said Mello. "But what modeling does is different from polling. We collect a huge amount of data about voters and then create a model of each voter by looking at hundreds of variables such as what groups they belong to, what magazines they subscribe to, education levels, etc.," said Mello. "We also did dial testing, which involves showing voters actual opposition ads and messaging, and then measuring their visceral reactions."
Mello explained that FFAA's modeling used data collected by Michigan's Democratic Party over the last decade in a huge database know as the Voter Activation Network. VANs are now used in almost every major democratic campaign in America.
"By using this sophisticated statistical modeling we assign each voter a score from zero to 100 indicating how they will vote on this specific issue in this specific election," said Mello. "We then do a prediction based on the collective electorate, crossing it with other models such as voter turnout expectations."
She said FFAA has found modeling to be a more accurate predictor of voter behavior, especially on LGBT issues. "When the opposition comes out with exaggerated and emotional claims that voters may hear for the first time, we found that polling was not capturing what happens when people actually vote on LGBT issues," said Mello. "It takes a lot to combat the negative impact of the opposition ads. It requires lots of inoculation ahead of time for voters to reject that messaging."
Using the same techniques, FFAA's modeling analysis predicted the Nov. 3 Houston vote, which struck down the city's nondiscrimination ordinance, within two percentage points. Other polls in Houston were far less accurate with some predicting a win.
"This is not a typical issue," said Mello. "It isn't necessarily true that young people, Democrats, women, are favorable on our issues. It is far more complicated than that."

Capacity – What Would It Take To Win?

If the FFAA model is correct and the electorate is only at a 42 percent YES vote, the next logical question is: Do we have the resources to move the number over 50 percent by next November?
Mello said that given the high voter turnout expected in Michigan next year – 62 percent in a presidential election – FFAA's model predicts the ballot initiative would lose by 720,000 votes. "That means we would have to change the minds of 365,000 voters. That is a daunting task," she said.
State Rep. Jon Hoadley, D-Kalamazoo, has consulted and advised on hundreds of campaigns across the country. He ran the successful campaign to elect Bridget McCormack to the Michigan Supreme Court in 2012. He is well versed in Michigan's VAN and how it has been used to do modeling in Michigan elections, including his own successful race last year. Hoadley is recognized nationally as a leading expert on field organizing, which is the hard work of making direct contact with voters by knocking on doors, phoning and sending direct mail.
"The size and scope of the field operation is determined by how many minds have to be changed. First you have to find the voters who are persuadable and that are likely to vote," he said.
"The type of conversations that we are talking about are not casual conversations with your friends. Obviously we hang out with people that have a similar clustering of values and opinions as our own. But we will want to find those unlikely conversations and have long-form conversations about LGBT rights with people who are unfamiliar and/or hostile to the idea," said Hoadley. "This is arduous and takes a long time because we're dealing with fears and prejudices."
Hoadley guessed that to move 365,000 voters to a YES vote could require about 100 field organizers working full time between now and the election.
"It is sobering but has edges of hope. Even with all the success on marriage equality in the last decade, we have a long way to go before we reach full LGBT equality. That said – there is a path forward," said Hoadley. "Ten years ago the numbers were very similar on marriage equality. In 2004 the marriage vote in Michigan was 41 percent. When we take the time to do the education we can change hearts and minds. Given the marriage work that was done, maybe it will take less time. But the gap right now is huge and we need to figure out how to close it."
Stephanie White, executive director of Equality Michigan, was political director for the "Schauer for Governor" campaign in 2014 and the Michigan Democratic Party. After seeing the modeling numbers from FFAA, White became concerned that a ballot initiative would take far more resources – volunteers, time, energy and money – than the ballot question committee leaders expect. White invited Mello to attend the Dec. 7 forum to present FFAA's findings.
"If you have the voters to win and the framework is settled in people's minds that this is an issue of fairness, then the opposition has a harder job to change the framework," said White. "We have neither now. The press is already talking about this in terms of religious freedom."
White, who does not believe there is a way to win a ballot initiative in 2016, said that to change 365,000 votes we need to talk to 3.6 million voters. "You have to have 3.6 million individual conversations and really about six communication touches per voter to move them. That's a total of 22 million touches – mail, phone, visit, etc.," said White. "We simply do not have that capacity."

Who is Freedom for All Americans?
Freedom for All Americans is the national campaign dedicated to ensuring that every American, regardless of where they live, is protected under the law from discrimination on the basis of sexual orientation and gender identity & expression – without allowing overly broad and harmful religious exemptions that will encourage employers, business owners or others to choose to disregard those protections. With the ultimate goal of securing federal statutory protections for LGBT Americans, FFAA works at the federal, state and local level to advance nondiscrimination measures and laws.
FFAA's programs include the LGBT University, a training program to develop and support current and future nondiscrimination campaign leaders and staff. LGBT-U's nationwide faculty and researchers include some of the LGBT movement's most experienced and sophisticated political operatives.
FFAA is headquartered in Washington D.C. and has a staff of 30 professionals located across the country including:
Executive Director Matt McTighe has managed campaigns for governor, U.S. Senate and Congress and ran or advised state and municipal ballot initiatives across the country on a range of issues. He previously served as the Marriage Project Coordinator for Gill Action.
Kasey Suffredini, chief program officer of FFAA, has helped to elect 116 pro-LGBT candidates to office, including the nation's first openly gay state attorney general. Prior to joining Freedom for All Americans, Suffredini was the executive director of MassEquality, the Bay State's statewide LGBTQ advocacy group.
Sarah Vaughan, state legislative consultant, was the policy director at Gill Action. Before joining Gill Action, Sarah served as political director at the Democratic Legislative Campaign Committee, working to retain and win democratic majorities in state Legislatures throughout the country. Sarah also worked as caucus director in the Colorado senate for the first female president of the Colorado Senate, Joan Fitz-Gerald.
Amy Mello, public engagement director, was the field director at Freedom to Marry from 2013 through 2015. She was field director for five statewide campaigns to win marriage for same-sex couples, including three years working to protect Massachusetts' first-in-the-nation marriage rights from legislative attempts to repeal the law or place it on the ballot for public referendum. In addition to Massachusetts, she has worked on marriage campaigns in Connecticut, California, Rhode Island and Maine, the country's first-ever popular vote victory for marriage.