Do We Really Know Anything About Changing Minds?

A chart from LaCour and Green’s study of voter attitudes before and after they spoke with either a gay or straight canvasser about their views on same-sex marriage. These data are now disputed. (Source: Mike LaCour’s web site, accessed May 27, 2015.)
By Michael Kuczkowski

“… just because the data don’t exist to demonstrate the effectiveness of this method of changing minds, doesn’t mean the hypothesis is false.”

This sounds like a weird riff on a famous line from Joseph Heller’s Catch-22. But, this isn’t a parody, and it’s not really funny.

The quote is from Columbia University Prof. Donald P. Green, a leading political scientist. He said it while talking to This American Life last week, after revelations that a widely heralded Science magazine article he co-authored with a graduate student from the University of California at Los Angeles may have been based on data fabricated by that graduate student.

If you read our blog, you probably know about the study. In it, gay canvassers spoke to Los Angeles voters who opposed same-sex marriage, asking them about their attitudes on the issue. After a single 20-minute conversation with the canvasser, who revealed his or her sexual orientation in the process, those voters who initially reported being against same-sex marriage said they now supported it. A year later, according to the study, they still did.

The study was a national sensation. The New York Times, The Washington Post, The Wall Street Journal wrote about it. This American Life devoted an entire episode to the study, and has now posted blogs (here and here) explaining what went wrong.

Last week, Professor Green sent a letter to Science asking that they retract the article; the publisher says the issue is under review. Michael J. LaCour, the graduate student in question, stands by his study, so far. His position appears shaky. He says he will issue single, definitive response by Friday, May 29.

Here’s what happened, according to news reports and a review of a critique of the study published online by a pair of researchers at Berkeley: Those researchers – graduate students Joshua Kalla and David Broockman – were intrigued by LaCour’s results. They tried to duplicate his research methods in a follow-up study, and they failed.

The more questions they asked, the more the study began to collapse like a house of cards. Confronted by his graduate advisor at UCLA, LaCour admitted that he lied about a significant amount of grant money that he claimed he was using to pay research participants. He changed his story about the incentives he used, saying that instead of paying participants $5 per person, he offered them a chance to win an iPad. Then he said he’d accidentally deleted his data set, which had been publicly available (and was accessed by Kalla and Broockman).

Meanwhile, as Kalla and Broockman dug in further, they found a number of statistical anomalies. LaCour’s data seemed almost identical to data from a 2012 national data set – even though his data set was of Los Angeles voters (hardly reflective of the nation). Lacour said he used a kind of sampling technique that would be highly unlikely to replicate any other data set, so the patterns didn’t make sense. His data were too perfect, replicating the other study’s distribution patterns and showing no outliers typical of large-scale quantitative research.

Suddenly the remarkable results seemed remarkable for an entirely different reason: A far-reaching allegation of fraud.

The questions about the study call into question a number of issues about academic research, graduate student supervision, peer-reviewed publications and mainstream reporting. And, they may well cost LaCour his job (he’d recently been appointed to a teaching post at Princeton.)

But, those are not really our concerns. We’re communicators. We want to know, is the science wrong? (And, on a more personal level, was my blog post last month wrong?)

The answer to those questions would appear to be no. But it requires some explaining.

Let’s say, as seems likely, that the piece by LaCour and Green is a fiction. If so, that means LaCour went to tremendous lengths to create a fictional data cohort that created the appearance of a credible study. He duped some pretty impressive folks and sent hundreds of canvassers out on what appears to have been a wild goose chase.

But, some elements of the study were real. The Los Angeles LGBT Center canvassers did go out and interview real Los Angeles voters. The audio clips of interviews that aired in This American Life were also real. And, some of the people, as interviewed, did change their minds as a result of that outreach, at least in the short term.

The statistical correlation between the sexual orientation of the canvassers and the degree and robustness of mind change appears to be completely fabricated. We also don’t know if people’s minds remained changed after that initial conversation.

But, the technique the canvassers used, which is what I focused on in my blog post, was not discredited. We don’t have any data to say whether it was effective or not. Which is a shame. As Green told TAL: “All that effort that went in to confecting the data, you could’ve gotten the data.” But they didn’t.

So what do we know about changing minds?

As the original TAL show said, it’s rare. Hard to predict, harder to control. The research we have on changing minds touches on myriad disparate fields, including mass communications, political science, cognitive behavior, economics, psychology, neuroscience, social networks, linguistics and sociology. There’s no silver bullet, but there are some important patterns.

Here are three sources that provide credible guidance on the subject:

Howard Gardner’s 2004 book Changing Minds offers up insights on what causes people to change their minds in different circumstances and contexts. Gardner, a Harvard psychologist, highlights 7 “R’s” that are relevant to the process of changing minds: reason, research, resonance, representative redescriptions, rewards, real-world events and resistances (that must be overcome). He also highlights how different the process of changing minds is in different contexts: large-scale heterogenous groups versus large-scale homogenous groups; change inspired by art or science; formal, instructional mind change, such as schools; small groups or families; and one’s own mind. These distinct realms call for different approaches.

John Kotter has been talking about change within organizations for decades. His book Leading Change, outlines an eight-step process for leading change, aimed at organizational leaders. While his process is focused on organizational change, communications plays a vital role in most steps. (Interestingly, Kotter’s model is built on his analysis failed organizational change efforts, rather than successful ones.)

Finally, this year the Institute for Public Relations recently launched a Behavioral Communications research program, led by Dr. Terence Flynn and Ogilvy’s Christopher Graves, to examine the role of communications in driving behavioral and attitude change. The results of that team’s extensive literature review are consistent with many of the precepts we’ve touted here: the messenger matters; emotional messages trump analytical ones; the “backfire effect” (in which people harden their beliefs in the face of contrary evidence) is real; and narratives are powerful vehicles for persuasion.

Flynn presented a draft of their model for changing minds at a conference I attended in January, and I think it’s very interesting. Watch that space, I’m a believer in their science.


Screen Shot 2015-05-27 at 10.19.55 PM
(Source: Pew Forum)

As for the issue of same-sex marriage, we know that Americans have, in fact, changed their minds. Polls by the Pew Research Center for Religion & Public Life, for example, show that 52 percent of Americans support same-sex marriage today, up from 35 percent in 2001.

What we don’t have is a large-scale experimental design study that tests the theories above in a “live” environment, which would be great to have, without question.

What’s interesting about this is that LaCour may well benefit from a kind of reverse “backfire effect.” LaCour’s study was really a narrative focused on the connection between the sexual orientation of the canvassers and the lasting effects of changed minds. Same-sex marriage is, for many people, an emotional and deeply personal issue. It all adds up to a remarkable story, even if the data were bogus.

I wonder if, as with the discredited research on a link between autism and vaccines, people who initially found the study compelling will forget it was discredited and continue to believe that gay canvassers were able to change minds in a real and lasting manner.

Still, something about it seemed too easy. It suggested an arms race of mind-changing outreach campaigns: If only we could mobilize the representatives of a disenfranchised group to engage those who supported oppressive policies, we could change the world. Then again, so could our opponents.

Meanwhile, we’ll continue to focus on the tools and techniques we know work: credible messengers, strong narratives, emotional connections and an awareness of the environment. It’s not easy, but it’s what makes for very interesting work.

How Not to Handle a Crisis: A DeflateGate Update

Tom Brady, who is implicated in an NFL investigation into whether his team intentionally deflated footballs used in the AFC Championship Game in January. Photo by Keith Allison. Used under Creative Commons license.

By Mike Kuczkowski

In Week 7 of what was a tough 2014 NFL season, Patriots Quarterback Tom Brady was ticked off.

It was halftime, October 16. Patriots v. Jets. At home. New England was up 17-12, in a surprisingly feisty matchup. At the time, the Patriots were 4-2, having endured two tough losses in the first four games of the season. Brady had two touchdown passes in the game already, but he was annoyed at the condition of the footballs he was throwing.

He turned to John Jastremski, Patriots equipment assistant, and told him the footballs “f***g suck.”

Whether that was a flippant comment or the start of a conspiratorial effort to ensure that Patriots footballs were pressurized to below regulations, we still don’t know. We do know that four weeks later, against the Indianapolis Colts, the Colts intercepted a football from Brady that felt “squishy” in their estimation. They tucked that little insight into their proverbial back pocket, waiting to call the Patriots out on it if it ever became useful. (I’ve said it before, and I’ll say it again: this is the NFL’s Pine Tar incident.)

Fast forward to the AFC Championship Game, January 18. Patriots v. Colts again. Second Quarter, Patriots ahead 14-0. Colts linebacker D’Qwell Jackson intercepts a deep pass up the middle intended for Patriots tight end Rob Gronkowski. Jackson hands the ball over to Colts team officials, who test the ball and find it’s underinflated. They flag it to the NFL, alleging foul play.

And so, Deflategate is born.

This week, the scandal trudges on, with the release of a 243-page report from the law firm the NFL retained to investigate the Colts’ allegations. The DeflateGate crisis erupted the next day, dominated the airwaves in the two weeks before Super Bowl XLIX, and now, it’s back.

It turns out, not surprisingly, there are employees of football teams who spend a lot of time preparing footballs. Breaking in the leather, reducing or increasing the tackiness of the ball. Air pressure, at least for some, is actually a pretty big issue. Clubhouse attendants joke a lot about the fussy demands of quarterbacks. Signed game day jerseys and footballs are given out in gratitude to the men behind the scenes. It’s a big deal.

And, it’s not without drama. Among other disclosures, we can now read the text message exchanges between Jim McNally, Patriots locker room attendant, and Jastremski – the two men responsible for delivering game day footballs that Quarterback Brady would find acceptable – that are as ribald as any of the jokes about Brady’s “balls” published in the midst of the scandal.

In one exchange, October 17 – the day after the Jets game – McNally tells Jastremski that he plans to overinflate footballs, just to get back at Brady for complaining.

“Tom sucks…im going (to) make that next ball a f**in balloon.”

Whether these texts represent a “smoking gun” of conspiracy, or innocent banter between two colleagues is up for debate. (Check Twitter, the debate is happening.) But, it further cements DeflateGate as one of the more ineptly handled crises of recent years.

I opined on Deflategate back in January, and said there were a number of clear missteps by the Patriots and the NFL in their handling of the crisis. In summary, I said the Patriots (and to some degree the NFL):

  • Bungled their disclosure of facts
  • Fumbled the roll-out of information about the incident
  • Offered up extremely amateurish spokespersons
  • Failed to manage the calendar and resolve issues in a timely manner
  • Failed to provide context for the importance of the issue at hand

Let’s add “waiting until early May to release a definitive account of what actually happened” to the list. And, that goes onto the NFL’s ledger. Conveniently after the NFL Draft. They appear to have managed the calendar a bit, shall we say, too deftly.

That said, the report itself is revealing in its facts. Startling in its candor. Along the way, it raises a host of issues that spell trouble for the reigning Super Bowl Champions. Specifically:

  • Jastremski and Brady talked a lot about the condition of game day footballs, and Brady took a personal interest in virtually every aspect of how those footballs were prepared
  • Jastremski and McNally joked a lot about that process
  • McNally, who was responsible for taking the game balls that the referees had approved onto the field, left the officials’ locker room with the footballs without permission, which is a breach of standard operating procedure (Walt Anderson, the head of the officiating crew, said it was the first time in his 19 years as an NFL official that he could not locate the game balls at the start of a game.)
  • Game officials did not accompany McNally and the balls to the field, as is standard practice
  • McNally stopped along the way and took the balls into a bathroom and locked the door for 1 minute and 40 seconds

If any deflating was done, that’s when it happened. (Incidentally, McNally’s nickname, according to the report, was “The Deflator.”)

So the whole DeflateGate controversy probably was not much ado about nothing after all.

Perhaps most significantly, it makes clear that at least at some points in his career, Tom Brady has been very concerned about the issue of the amount of air pressure of his game day footballs.

And as early as 2006, according to reports, Brady was one of several quarterbacks who lobbied for a rules change that would allow visiting teams to have more autonomy in the preparation of game day footballs. In the report, as noted above, he definitely gives staff clear instructions on the condition and the pressure he wants in game day footballs.

The report also explains something I criticized back in January – and suggests a different explanation for the dynamic I observed than I expected.

In head coach Bill Belichick’s Jan. 22 press conference, he said he was “shocked” about news reports of deflated footballs and had no knowledge about the process of preparing game day footballs. He was monotone and came across curmudgeonly, as he often does. Belichick ended it by saying the media should ask Brady about the issue: “Tom’s personal preferences on his footballs are something he can talk about in much better detail and information than I could possibly provide.”

Belichick’s claims of ignorance ran against his reputation as an incredibly detail-oriented control freak. And, his “shocked” quote drew comparisons to Captain Renault’s declaration of innocence about the gambling at Rick’s Café Americain in Casablanca.

A few hours later, Brady projected innocence and even bewilderment at the issue. In response to the question “Is Tom Brady a cheater?” Brady replied, “I don’t believe so.” Asked if he knew whether anyone on the Patriots had done anything wrong, he said “I have no knowledge of any wrongdoing.” In addressing how he liked footballs prepared for game day, he said he liked them inflated to 12.5 psi, which is the lowest level permitted by the rules, but that his process in general focused on picking gameday footballs based on their grip, not their inflation level. “It’s not like I ever squeeze the football, I just grip the football.” To many observers, Brady saved the day.

So, yay for Brady and boo (again) for Belichick. Except, the NFL report concludes in fact Belichick did not know anything about the issue. He was being honest and, I think, fairly transparent. Now it looks like Brady is at least stretching the truth, according the facts outlined in the report.

What’s next? Sadly, this report – 14 weeks in the making – doesn’t bring us what we need: Closure.

The report is detailed on the facts, yet equivocal in its conclusions: “(I)t is more probable than not that New England Patriots personnel participated in violations of the Playing Rules and were involved in a deliberate effort to circumvent the rules.”

Now what?

That shouldn’t even be a question. There is no reason this report should have been released without at the same time issuing suspensions or fines. It should be a one-day story. Rip off the Bandaid. Deal with the consequences. Move on.

Instead, it’s as though the report is a trial balloon, intended to gauge the public’s reaction to the facts before the league makes its move.

This shows the biggest problem the NFL has, the lack of a disciplinary structure for issues of this nature. Commissioner Roger Goodell lacks trust, and candidly as the head of the organization he shouldn’t be the one doling out discipline. Someone else, someone credible, even-tempered and with a deep reverence for the integrity of the game needs to have the role of fining and suspending players and teams that violate the rules.

Goodell will eventually address DeflateGate. He may suspend Brady for a game or two, banned the clubhouse employees from the league or fine the team. For now, all we can do is speculate.

Which means more debate on an issue that should already be in the rear-view mirror. Meanwhile, the list of ways in which DeflateGate is a case study in how not to handle a crisis grows longer.

Listening, Gay Marriage and Lessons for Changing Minds

A pro-gay marriage rally in Sacramento in 2008. A recent study showed that engaging opponents to gay marriage in a dialogue about the issue was effective in changing minds. Photo by Kelly B. Huston. Used under Creative Commons license.
By Michael Kuczkowski

Years ago, I was part of a team that was hired to help a pharmaceutical company explain to America why its industry was a good thing.

This was not an easy task, obviously. The pharmaceutical industry has a weak reputation overall, despite delivering many innovations that improve the health of people around the world. A big chunk of the problem is rooted in costs: People believe health care is a right, yet prescription medications cost money, sometimes a lot of money. With ever-changing insurance co-pays, drug classes and deductibles, individuals increasingly bear the brunt of those costs. Which is tough.

It was, however, a fight I could feel good about. The industry was taking a lot of criticism for things that I didn’t think actually were bad things. And there were lots of ways in which the pharmaceutical industry was (and is) hugely valuable. The prescription drugs these companies produced often saved lives, and certainly extended the lives of many, many people. It didn’t make sense to me that the industry was being demonized as much as it was.

The client in question had amassed mountains of compelling economic and policy arguments – all of which I think have merit – about the value of the innovative pharmaceutical industry. (I know this remains a controversial question in many quarters, but that’s an argument for another time.)

What became clear early in the engagement was that the leaders of this company felt strongly that if only they could get people to understand these facts – very credible analysis by respectable academics – then people would come to love the pharmaceutical industry, and the industry’s reputational problems would go away.

Unfortunately, that belief was unrealistic. Few things are as emotionally charged as health care, and it’s the emotional vector of rising health care costs, rooted in very real experiences, that shapes people’s opinion of industry. (There are other factors too, but this is the single biggest driver.)

We came up with a three-part framework for dialogue.

Understanding: You have to start by listening and acknowledging facts that no one can really dispute: health care is expensive, more and more people are bearing the direct costs of health care, people are suffering, the system is maddeningly complex and changing in ways that are painful to the average American. Expressing empathy for what people were going through in their battles with the health care system is critical. And it’s not hard. Listen to people’s stories. Tell them you understand their frustration. And recognize that their frustration is real.

Facts: Use the facts, if you must, to illustrate the value the industry brings to people and the health care system. Innovation saves lives, and no other industry invests as much of their revenue in research and development as pharma does. For example, while new drugs are costly, they also reduce other types of health care spending, often by a significant margin (the linked study showed that new drugs increased pharmaceutical spending by $18 per person, but reduced other types of health care spending by $129 per person. That’s a big-picture “good thing” that most people don’t know about, and could be useful in reframing their thoughts about the value of the industry.)

Future: Focus on a future where more people have access to medicines, medicines cure more diseases and prolong life, and the cost burden borne by individuals is more aligned with the things that support prevention and interventions that delay disease progression. After all, it is possible to change policy around health care financing to support things, like regular doctor visits, disease management and access to medicines, that can keep people healthy and reduce their individual cost burden. We just have to fight for it.

So there it is. U-F-F. UFF. Unlikely to win a Cannes Lion for best mnemonic device, but more effective than trying to give the entire U.S. population an advanced degree in health economics.

UFF didn’t go very far, ultimately. We tried it, and we trained people on it, but a leadership change prompted a shift in the overall approach. None of which, in hindsight, worked out very well.

Still, I was reminded of it recently while listening to a podcast from This American Life titled “The Incredible Rarity of Changing Your Mind.”

A big chunk of this program examined the experience of a group of people who canvassed voters last year about California’s Prop 8. You may have read about this based on a study published in February in Science magazine: Researchers studied an effort by canvassers to persuade voters to change their minds from neutral or opposed to gay marriage, to supportive of gay marriage. The canvassers, many of whom were gay, were actually incredibly successful in getting people to change their minds. Follow-up research showed that people’s changed minds stayed changed for up to a year. Both the degree of changed minds and the length of the effects were a surprise to experts.

The reporting on this focused on the aspect of the findings that I just emphasized: Gay canvassers were more effective at changing minds on the issue of gay marriage than straight ones, statistically speaking. In other words, the messenger matters. (And in communications, we know this is true.)

But there was more to it than that. It turns out that the approach the canvassers adopted matters too.

First, they listened. Rather than making some compelling argument about fairness or the rights of individuals, canvassers got the voters they were interviewing to talk about their own personal experiences with gay people or marriage. And, they talked about whether being gay was a choice or a genetic predisposition. Ultimately, the discussions led somewhere. Like, wouldn’t you want the gay people you know to enjoy all the benefits of marriage?

As I’m describing this, the second point is pretty obvious: It’s an emotional conversation. We’re not talking about facts. There were no data points about how many gay people can’t visit their partners in the hospital because they don’t get spousal recognition, or how many people are denied health care because they don’t have coverage from their lover’s health plan. Which would be valid arguments to make, of course. Instead, the conversations focused on what it felt like to know someone who was gay, to feel what they must feel, to miss out on the things a heterosexual married couple naturally enjoy by dint of being married.

The third thing was how artfully the canvassers brought the voters to their own realizations. They would weave in the fact that they were gay at just the right moment. Or, even just wait while the voter connect some key dots. (One interviewee, a widower whom the broadcast named “Mustang Man,” completely flipped his stance in the course of the interview with a canvasser, because he reflected on the powerful love he’d had with his wife: “I would hope that they would find the happiness that I had with my wife. If you could have that kind of relationship with your partner or the other sex, I would say you’re a very lucky person. Because I know I had it. But yeah, that’s what I would wish on them. That they’d be as happy as I was with mine.”) It was subtle but impactful, holding a kind of mirror of reality up to people’s faces.

And that’s how the interactions worked. Canvassers asked voters whether they were likely to vote for gay marriage before and after the discussion, and many of them were more likely at the end of the conversation. By statistically significant margins. And, the opinion change held for a year in many cases.

There are incredibly powerful lessons in here for communicators.

Firstly, if you want to change minds, you have to start by listening. This is huge. It’s almost the opposite of the job, right? I’m a communicator, therefore I should talk all the time. But, we know that understanding where people are is critical to helping them adopt a new opinion or behavior.

Second, you can’t change someone else’s mind for them. This seems obvious, but the process has to be owned by the person who is changing their mind. Facts are not going to help you (oh, how I wish I could travel back in time and tell my mid-30s self and pharma industry clients this truth). In fact, studies show that if people with a certain opinion are confronted with facts that show they are wrong, those people are more likely to become even more entrenched in their mistaken beliefs than before. (This is known as the “backfire effect”) But, you can lead someone to reframe their thinking on an issue, and it can even happen quickly. (This is where I’d go back and change the middle “F” in the UFF framework above to say, try to find ways to help the person you’re talking with connect their own dots about an issue, rather than laying a white paper on them.)

Finally, and most importantly, the emotional aspect of these arguments is incredibly significant. People feel hurt, anger, joy, outrage, frustration, anxiety and all kinds of other emotions about their experiences. Mounting factual arguments against a wall of emotions is a waste of time and energy. It’s not going to change minds.

But minds can be changed. The human mind can learn and evolve and take up new opinions over time.

This has huge implications. It suggests that organizational change efforts should think not just about who occupies the newest boxes on an org chart, but also about the emotional state of people across an organization.

It reinforces the vital importance of research in understanding how people view issues.

It shows how important it is to listen, be flexible and adapt in the face of unanticipated data.

And, it shows that communications can be incredibly effective – but that it’s also very resource intensive. Changing minds on the gay marriage question wasn’t going to happen from exposure to a few 30-second ad spots. It was going to take a dialogue, and an artful one at that.

But that’s okay. Because the issues on which we need to change minds – racism, police brutality, gender equality and even the divisive issue of health care– are important issues. They’re worth the time. Because on those and many other issues, every opinion counts.