Preach To More Than the Choir

MAGA supporters sang the national anthem during AmericaFest, the first Turning Point USA summit since the death of Charlie Kirk, in Phoenix, Arizona, on December 20, 2025. CHENEY ORR / REUTERS

Note from the author: This is a long article. That’s because it explores a complex question in-depth and with research backed solution oriented journalism. Enjoy and thanks for being here.

What’s the reason I’m writing this article? Because there’s a question hanging over all of our heads. How do we get out of this mess? Put another way, what separates successful change movements from unsuccessful ones? That question has many legitimate answers and none of them are simultaneously simple and easy. This article looks for evidence supporting one of those answers, the one about the preaching to the choir problem and the modern propaganda bubbles that make it harder to solve.

So when looking for answers, I often start by asking who does it best, or most successfully. In this case I didn’t have to go looking, because I’ve spent years researching and writing on this, and the answer is intelligence agencies and spies. Throughout history we can find examples of the most severe propaganda bubbles being targeted, causing defections and unrest, through both covert and overt methods.

There are many types of propaganda bubbles in the modern information environment, and inside any one of them you find several types of people, each responding to the same message differently. Some have started to doubt what they’re told but haven’t said so to anyone. Some have already checked out, going along out of habit, halfway to leaving. Some are in an apolitical bubble, tuned into other things and picking no side. Some would already be with you if they had any idea you existed. And some still believe all of it, but would act differently if the conditions were right.

The defector research points to a clear pattern in why people leave. They left mainly because of personal disillusionment, and only rarely because they had reasoned their way out of the regime’s ideas. Wilhelm Marbes, a CIA psychiatrist, spent years studying defectors and ranked political conviction among the least important motives. He stated the finding about as plainly as you can: nobody ever defected because he was happy. The cause was personal. The regime broke a promise, or the person’s life stopped making sense.

Personal disappointment works only when the person also understands what caused it. This is what effective propaganda is built to prevent, by making sure people blame someone other than the regime. The citizen whose life got worse under Stalin is taught to blame America, and as long as he believes that, he never connects his disappointment to the regime that caused it. Defection comes only when the person identifies the real cause, which is the connection an operator has to make possible.

This points to a distinction that researchers who study how people leave extremist groups treat as fundamental. Stopping the behavior and changing the belief are two separate things, and one can happen without the other. Researchers call the first one disengagement, when a person stops acting on the belief, stops showing up, stops doing the work. They call the second deradicalization, when the person actually stops believing.

John Horgan, who interviewed people who had left terrorist movements, found that behavior is easier to change than belief. A study of about 300 former American extremists found that roughly a third had changed on all counts, behavior and criminal activity and belief together, while about a fifth had stopped the criminal activity but still held the beliefs and kept the associations. Changing the belief was the exception. Stopping the behavior was the more common result, and it usually came first.

As counterintuitive as it sounds, it isn’t very difficult to change someone’s behavior. You can get a person to act in a way they normally wouldn’t while they believe they are acting of their own free will. And self identity, even ideology, can follow to match the behavioral change.

The free-will piece, that people act differently while believing the choice was their own, is well evidenced. In Festinger and Carlsmith’s 1959 induced-compliance study, people paid one dollar to tell a lie ended up believing it more than people paid twenty dollars, because the twenty-dollar group had an external reason for lying and the one-dollar group didn’t, so they resolved the discomfort by changing their actual attitude. Small external pressure produces real internal change, and the person experiences it as their own conclusion. Bem’s self-perception theory describes the same thing from another angle: people infer their own attitudes by observing their behavior, the same way an outsider would. You find out what you think by watching what you do. Under all of it is cognitive dissonance, Festinger 1957, the discomfort of holding a belief while acting against it, resolved by changing the belief.

The foot-in-the-door effect, Freedman and Fraser 1966, is the same mechanism used as a deliberate technique. People who agreed to a tiny request later agreed to a much larger one they would have refused outright, because the first small action changed how they saw themselves. Disengagement, foot-in-the-door, and induced compliance are the same mechanism at different scales. Get the behavior first, by small steps and low pressure, and the person revises the belief to justify what they have already done, and experiences the new belief as their own.

One warning belongs here first. Some of these methods are manipulation. They change what a person does without their knowledge or consent. A person’s knowledge and consent are both necessary prerequisites for using and developing these techniques ethically. We study these methods anyway for two reasons. You cannot resist what you do not understand, and the people inside a propaganda bubble are targets of these methods every day. And some of these methods can be separated from the manipulation and used honestly, which is part of what this article is looking for.

The intelligence and military services treated persuasion as a practical discipline, something you apply in the field and test against what actually happens. Their methods were based on what psychologists understood about why people change their minds. Paul Linebarger, an Army colonel who helped organize the country’s first psychological warfare section, wrote the doctrine the US Army used for decades. His doctrine was based on audience analysis, studying the specific people a message is aimed at before writing a word of it. Daniel Lerner, a social scientist who served in the OSS, the wartime predecessor of the CIA, wrote a detailed account of the Allied campaign against Nazi Germany, drawing on his own work inside it. He devoted part of it to whether any of it actually worked.

The clearest example is the surrender pass. The Allied teams designed a leaflet that a German soldier could carry to Allied positions and use to surrender, and they built it from what they knew about their audience. German soldiers responded to official documents, so the pass was made to look official, with formal language and the look of something issued by an authority. Lerner judged it the most effective leaflet of the war. It was dropped at more than ten million copies a month, and the Allied command thought it worked well enough that in 1944 it forbade reproducing the design on any other leaflet. But Lerner reported the harder finding alongside the success. Most leaflets, he concluded, showed little evidence of dramatic effect, with the surrender appeals working best when defeat was already close. The assessment counted the failures, not only the successes.

The next problem was a whole population held inside a closed information system. A regime that controls the newspapers and the airwaves can decide what tens of millions of people know, and the defection research already showed why that matters, because a person who never learns the real cause of his disappointment never acts on it. The American answer was to broadcast the news in from outside the country. Radio Free Europe and Radio Liberty sent uncensored news into the Soviet bloc in people’s own languages, and the audience was large. Researchers later estimated that during the 1980s a quarter of the adult population in the bloc was listening, and that roughly a third of Soviet urban adults heard Western radio. Measuring that audience took real work, because you cannot run a survey in a country that would arrest the respondents. The researchers estimated it from listener mail, from interviews with travelers and emigrants conducted in the West, and from computer models built to estimate the parts they could not measure directly.

The clearest case of what this could do is Józef Światło. He was a senior officer in the Polish secret police who ran interrogations and knew the system from the inside, and in 1953 he defected to the West. Radio Free Europe put him on the air. Across 1954 and 1955, in the broadcasts known as Operation Spotlight, he described, in detail and over many broadcasts, how the secret police actually operated, the torture, the surveillance, the corruption of the party elite, and the Polish service answering to the Soviet government. The broadcasts reached the country he had left, and the effect was concrete. Poland dissolved the Ministry of Public Security at the end of 1954. The ministry was reorganized, its staff cut sharply, its informer network reduced, and several of its top officials arrested and later sentenced to long prison terms. He left first, not from any argument, and his testimony, broadcast to the people still inside, produced the result.

The historical work was expensive and secret, aimed at a population the operators would never meet. The modern version is cheap and open, and better tested than any of it. It is called inoculation. You can make a person resistant to a manipulative message before they ever encounter it, by showing them a weak version of it first and explaining how the technique works. William McGuire established the basic effect in the 1960s. Afterward the person recognizes the technique when they later see it used for real. Two researchers at Cambridge, Sander van der Linden and Jon Roozenbeek, developed this into short videos and tested it at scale. Each video teaches people to spot one manipulation technique, like emotional language meant to produce a reaction rather than a reasoned response, false choices that conceal the real options, scapegoating, and attacking the person instead of the argument.

In seven preregistered studies, six controlled experiments with about 6,500 people and one field experiment on YouTube with more than 22,000, the videos measurably improved people’s ability to recognize those techniques. The YouTube test ran the videos as ads. The videos were shown to about five million users at roughly five cents a view. This works on a divided audience for one reason. The videos make no argument about which side is right. They teach a person to recognize a manipulation technique, so the effect was the same for liberals and conservatives. The limit is that it does not last. Within days or weeks the recognition declines, and people need a reminder to keep it.

Inoculation protects a person before they encounter the lie. The harder case is the person who already believes it. For years the fear was that correcting a false belief could strengthen it, an idea called the backfire effect, where telling someone the truth made them hold the falsehood more firmly. Later research found this is rare. Thomas Wood and Ethan Porter tested 52 contested issues with more than 10,000 people and could not reproduce it, finding that people generally accepted the accurate information when given it. The original worry was largely wrong, which means correction is worth attempting.

What correction it takes is specific. Stating the truth once does not undo a false belief a person has held and used, because the falsehood keeps influencing their thinking even after they accept the correction, an effect researchers call continued influence. The practice that works against it follows a clear sequence. You lead with the truth and state it plainly and first. You warn that a falsehood is coming before you repeat it, and you repeat it only once, because every repetition makes it more familiar and familiarity makes it feel true. You explain the technique that made the falsehood convincing, so the person sees how they were misled. Then you state the truth again, and you give the person a true account to use in place of the false one, because a correction that takes the false belief away without giving the person something true to replace it will fail, and the person will keep relying on the falsehood.

Inoculation and correction both assume a person who is open to getting things right. The committed believer is a different problem. For someone whose identity depends on a belief, the facts matter less than what the belief means for their standing in the group. Abandoning it would mean losing that standing. Dan Kahan, a law professor at Yale who studies how people reason about disputed facts, documented this. People do not evaluate a politically charged claim on the evidence. They evaluate it on what believing it would mean for their place in their group, and they accept or reject the evidence accordingly. Kahan found that the people best at quantitative reasoning were the most polarized on politically charged questions, because their skill made them better at constructing reasons for the conclusion their side required. They reason well, and they aim it at the conclusion they have already settled on. With a committed believer, you reduce the threat to their identity before you introduce the evidence. One way is to argue in their own moral terms instead of yours. Robb Willer and Matthew Feinberg found that the same argument became far more persuasive when it was framed in the listener’s values rather than the speaker’s. Another way is to have the person first affirm something important to them that has nothing to do with the disputed question. Geoffrey Cohen and David Sherman found this left people less defensive and better able to consider evidence that would otherwise threaten them.

The method matters, and so does who delivers it and in what form. A direct argument invites the listener to argue back. A story does not, because a person follows a narrative instead of debating it, and is persuaded without raising the defenses a direct claim provokes. This is the basis of entertainment-education, where a serialized drama models a behavior change and audiences adopt it, a technique developed by Miguel Sabido at Televisa in the 1970s and studied since. The other half is the messenger. The person most able to help someone leave a closed belief system is someone who was inside it and left, what the exit programs call a former. A former has credibility a critic never will, because they believed what the listener believes and can describe leaving as someone who did it. The programs that help people leave extremist movements, EXIT in Germany and Sweden, Life After Hate in the United States, rely on these former members rather than outside experts.

Here is where the warning from the start applies most directly. Everything to this point can be done with the person’s knowledge, and most of it openly. What follows was built to deceive. During the Second World War the British ran black propaganda stations into Germany, and the purest of them pretended to come from inside Germany itself. Sefton Delmer, a journalist who had grown up there, created Gustav Siegfried Eins, a station fronted by a character called Der Chef, who sounded like a loyal Nazi disgusted with the party’s corruption but devoted to the nation. The station was a British operation from the first word. Delmer built its credibility with real news and material a German listener wanted, then inserted the content meant to demoralize, a method he summarized as cover, cover, dirt, cover, dirt. The true material was there to make the rest believable. German security took Der Chef for a real internal conspiracy and set its codebreakers on the fake ciphers he read out each night. This is propaganda at the far end, where the audience does not know who is speaking to them or why, and the deception is the point. It worked, and a movement that wants the public’s trust cannot use it, because the method and the goal are in direct conflict.

Some of these seem pretty horrible, right? The method itself is neither good nor bad. Whether it is ethical depends on how it is applied, and the question is consent. Informed consent is the fullest form of it, where the person is told exactly what will be done and agrees to it, which is the right bar for medicine and research. Below that, the person agrees to less. You know an advertisement is an advertisement, but you never agreed to the techniques working on you inside it, which is why marketing is the grey area. At the bottom there is no consent at all, where the person has no idea it is being done to them, which is manipulation.

So this is the answer this article went looking for. A movement grows only by winning over people who do not already agree, and that is the harder and more valuable half of the work. The intelligence services and the researchers who developed this work learned how people actually change, and they built their methods to give a person an honest reason and a way to act on it. The figures here come from studying defectors and committed extremists rather than the average person inside a propaganda bubble, so the mechanisms apply more widely than the exact numbers do. The same toolkit can grow a movement or run a fraud, and the difference is whether the people on the receiving end keep their knowledge and their consent. Win over the people who are not yet with you. Tell them the truth while you do it.

If you’re looking for a quick daily effective activist action: Go to Openstates.org, find your state house rep and ask them if they support Public Banking like the state of North Dakota has with the Bank of North Dakota. Let them know if they want more information you’d love to send it their way. You can send them the article hyperlinked in this sentence.

Additional note: there’s a socio-psychological phenomenon called the Bystander Effect, where people assume someone else will act, so no one does. I mention it because we have millions of readers, and The Existentialist Republic needs 10 new subscribers per article to keep the lights on. We have hit that on only 1 of the last 15 days. Please don’t let an activist journalism publication be the reason you skip a meal or miss rent. But if you value this work, we need you to keep this machine crafting tools for destroying fascism.