Most of us remember the specific, localized anxiety of the rubber eraser. Sitting in a silent classroom, staring at a multiple-choice Scantron sheet, you hover your pencil over a filled-in bubble. A sudden wave of doubt suggests the answer might be C, not B. The institutional wisdom handed down by generations of test-prep tutors echoes in your head. Trust your first instinct. Changing your answer will only hurt you. You leave the bubble alone.
The data says otherwise. When psychologists analyze the eraser marks on university exams, they find that the vast majority of answer revisions move from wrong to right. The first-instinct fallacy is a cognitive illusion. We remember the agony of changing a right answer to a wrong one because it haunts us, while the quiet success of correcting a mistake fades into the background. We are terrified of revision.
In Think Again, Adam Grant argues that this fear of revision extends far beyond the classroom. It infects our careers, our political discourse, and our organizational cultures. We live in an era of accelerating complexity, yet we cling to our initial assumptions with white-knuckled desperation. Grant proposes that the defining cognitive skill of the twenty-first century is the ability to unlearn and rethink. He builds a framework for how to achieve this, moving from individual mental habits to interpersonal persuasion, and finally to organizational design. If you want to understand why your uncle refuses to abandon his political conspiracy theories, or why your company’s executive team keeps funding a doomed product line, the mechanics are laid bare in these pages. Whether knowing the mechanics actually helps you alter them — that’s a different question, and Grant is more optimistic about it than he probably should be.
Why We’d Rather Be Wrong Than Uncertain
The foundation of the book rests on a framework originally developed by the psychologist Phil Tetlock. When we engage with information, we rarely process it objectively. We slip into one of three occupational mindsets.
When our sacred beliefs are threatened, we become Preachers, delivering sermons to protect our ideals. When we spot flaws in another’s reasoning, we become Prosecutors, marshaling arguments to win the case and defeat the opponent. When we need to win over an audience, we become Politicians, campaigning and lobbying for the approval of our tribe. In all three modes, the conclusion comes first. The reasoning just fills in behind it.
Grant suggests a fourth alternative. The Scientist.
To think like a scientist is to treat your opinions as hypotheses and your decisions as experiments. You search for reasons why you might be wrong. Grant uses the rise and fall of BlackBerry to illustrate the stakes. Mike Lazaridis was acting as a brilliant scientist when he hypothesized that people wanted a wireless device for email, testing and iterating until the BlackBerry dominated the market. When Apple introduced the iPhone, Lazaridis shifted into Preacher mode. He worshipped the physical keyboard. He prosecuted the idea of a touchscreen as an inefficient battery-drainer. He politicked with his existing enterprise user base, ignoring the shifting consumer winds. He refused to rethink his core product, and his company collapsed.
It is a clean, elegant narrative. Perhaps too clean. Grant presents the Scientist mindset as a cognitive choice, a pair of goggles you can simply choose to put on in the morning. Tetlock’s original 2002 paper in Psychological Review positioned the Preacher, Prosecutor, and Politician modes not as bad habits, but as deeply embedded evolutionary adaptations. We preach to signal tribal loyalty, which historically kept us from being exiled to the wilderness. We prosecute to enforce social norms. We politick to secure resources within a hierarchy.
The Scientist mindset is evolutionarily unnatural. It requires burning massive amounts of caloric energy to override our survival instincts. Daniel Kahneman, the godfather of behavioral economics, famously admitted that even after fifty years of studying cognitive biases, he still fell prey to them in his daily life. Grant knows this literature intimately, yet his relentless optimism sometimes glosses over the sheer biological friction required to maintain epistemic rationality. You can tell an executive to think like a scientist. When her annual bonus is tied to the success of a specific project, her inner Politician will almost always take the wheel.
The View from Mount Stupid
Grant spends significant time exploring the Dunning-Kruger effect, the psychological phenomenon where people with the lowest ability in a domain overestimate their competence the most.
They suffer from a dual burden. They are incompetent, and they lack the metacognitive skills to recognize their incompetence. They stand at the peak of what is affectionately termed Mount Stupid.
A little knowledge creates an illusion of explanatory depth. You read one article about macroeconomic policy and suddenly feel equipped to debate interest rates with a central banker. Grant contrasts this arrogant ignorance with its mirror opposite. Impostor syndrome.
Typically, impostor syndrome is treated as a pathology to be cured. We are told to banish our self-doubt and lean in. Grant introduces research by Basima Tewfik suggesting that impostor syndrome can actually be a performance enhancer. In a study of medical students, those who experienced frequent impostor thoughts did not make worse diagnoses. They actually performed better in their bedside manner. Because they doubted their own brilliance, they asked more questions, listened more carefully to the patients, and gathered better information — self-doubt pushing them, almost accidentally, into the Scientist’s chair.
This reframes confidence not as a measure of how much you know, but as a measure of how much you trust your ability to learn. Grant calls this confident humility. You can be confident in your ultimate capacity to solve a problem while remaining entirely humble about your current understanding of it.
The literature surrounding Dunning-Kruger has faced heavy scrutiny in recent years. Papers by researchers like Gilles Gignac and Marcin Zajenkowski suggest the effect might be less about a unique cognitive blind spot and more about a statistical artifact. Specifically, regression to the mean and the fact that absolute novices simply have more room to overestimate their scores than experts do. Grant acknowledges this debate briefly in the endnotes, but keeps the concept front and center in the main text. It makes sense why. Even if the statistical math of the original 1999 paper is debatable, the behavioral archetype is instantly recognizable to anyone who has ever logged onto the internet.
The Friction Fallacy
We generally view conflict as a negative force, something to be minimized or managed. Grant introduces a vital distinction from organizational psychology: the difference between relationship conflict and task conflict.
Relationship conflict is personal, emotional, and destructive. It is the friction of clashing egos and incompatible personalities. Task conflict is intellectual. It is the friction of competing ideas, strategies, and perspectives.
Tracking teams in Silicon Valley, researchers found that low-performing groups typically experienced high relationship conflict from the start. They disliked each other, which prevented them from doing good work. High-performing groups did not experience a total absence of conflict. They experienced high task conflict early in the project, which slowly tapered off as they aligned on a solution. They fought about the work, not about each other.
To cultivate this productive friction, Grant advocates for building challenge networks. We all have support networks — people who validate us and cheer us on. A challenge network is a group of trusted, highly disagreeable people whose specific role is to poke holes in your thinking. Grant uses the production of Pixar’s The Incredibles as the anchor case study. Director Brad Bird, facing an impossible budget and timeline, actively recruited the studio’s pirates. These were the disgruntled, disagreeable misfits who were frustrated with the company’s standard operating procedures. By empowering them to challenge the animation orthodoxy, the team delivered a masterpiece under budget.
I find this prescription conceptually brilliant but practically fragile. The idea that we can neatly separate task conflict from relationship conflict ignores the messy reality of human ego. Ideas do not exist in a vacuum. They are generated by people who bind their identities to their output. When you attack a designer’s prototype or a strategist’s financial model, you are attacking a piece of them.
In a sterile laboratory setting, participants can engage in pure task conflict. In a corporate environment, where promotions, status, and livelihoods are at stake, task conflict bleeds into relationship conflict almost instantly. Carsten De Dreu’s comprehensive 2003 meta-analysis on workplace conflict found that both relationship conflict and task conflict have strong negative correlations with team performance and satisfaction. Task conflict only avoids damaging the team when it is weakly correlated with relationship conflict — a state that is incredibly difficult to maintain outside of a controlled experiment.
Grant advises us to frame disagreements as debates rather than arguments, to use “how” questions rather than “why” questions. These are excellent tactics. Relying on highly disagreeable people to drive innovation often creates collateral damage that the literature rarely measures. The pirates might ship a great movie. They also might create a toxic culture that burns out the quieter, more agreeable talent in the room.
The Exhaustion of the Dance
How do you change the mind of someone who refuses to listen?
When we encounter someone with an opposing view, our default setting is the Prosecutor. We pile up facts, data, and logical proofs, attempting to bludgeon the other person into submission. This almost always triggers the righting reflex in our opponent. The harder we push, the harder they push back, digging in against everything we’ve offered.
Grant suggests we stop treating conversations as wars to be won and start treating them as dances to be choreographed.
The most compelling methodology he presents for this is motivational interviewing. Originally developed by clinical psychologists Bill Miller and Stephen Rollnick to treat addiction, motivational interviewing abandons the attempt to persuade. The interviewer asks open-ended questions, listens reflectively, and helps the person find their own internal motivation to change.
Grant profiles Arnaud Gagneur, a neonatologist in Quebec who uses motivational interviewing to address vaccine hesitancy in new mothers. Instead of lecturing parents with epidemiological data, Gagneur asks them about their fears. He acknowledges the confusing sea of information on the internet. He validates their desire to protect their children. By refusing to judge, he lowers their defenses. In his studies, a single motivational interviewing session in the maternity ward increased the likelihood of children being fully vaccinated two years later by nine percent.
People do not resist change. They resist being changed. That distinction — which sounds tidy on the page — carries enormous weight in practice.
We see this taken to its absolute extreme in the story of Daryl Davis, a Black musician who has spent decades befriending members of the Ku Klux Klan. Davis does not preach or prosecute. He asks them a simple question. How can you hate me when you don’t even know me? Over time, by listening and finding shared interests, Davis has watched over two hundred men voluntarily hand over their robes and leave the Klan.
It is an awe-inspiring story. It also reveals the heavy limitation of interpersonal rethinking. Motivational interviewing requires an immense expenditure of emotional labor. It demands endless patience, radical empathy, and the suppression of one’s own ego and outrage. It is entirely unfair to ask marginalized people to perform this labor for those who actively oppress them. Grant acknowledges this, noting that it shouldn’t have been Davis’s responsibility to educate white supremacists.
Even in less extreme scenarios, motivational interviewing is exhausting. It requires you to act as an unpaid therapist for your ideological opponents. The clinical literature on the practice is also more mixed than the book implies. A 2010 meta-analysis by Brad Lundahl and colleagues found that while motivational interviewing is effective compared to no treatment, it works less well when compared to other active treatments, and its effect sizes are generally small and diminish over time. Deploying these tools in the heat of the moment requires a level of emotional regulation that most humans simply do not possess when their values are under fire.
The Prism and the Pedestal
Peter T. Coleman runs the Difficult Conversations Lab at Columbia University, and his experimental design is simple and devious. He pairs up strangers who hold opposing views on highly charged topics like abortion. Before they talk, he gives them an article to read about a different divisive issue, like gun control. If the article presents the gun control debate as a binary — two extreme sides locked in battle — the subsequent conversation about abortion usually devolves into a shouting match. If the article treats the issue as a problem with many dimensions, competing values, shades of gray, the participants are significantly more likely to find common ground on abortion.
Binary bias is our tendency to seek clarity by dividing a spectrum into two mutually exclusive categories. The media feeds this bias by pitting climate change activists against climate deniers, ignoring the vast middle ground of skeptics, the cautious, and the disengaged. Grant argues that we must actively inject complexity into our public discourse. Acknowledging caveats and contingencies does not weaken an argument — it signals intellectual honesty and invites the audience to meet you halfway.
This principle extends into how we educate the next generation. Grant visits the classroom of Ron Berger, a teacher who treats knowledge not as static facts to be memorized, but as rough drafts to be revised. Berger’s students don’t just read history textbooks. They critique them for omitted perspectives and rewrite the chapters themselves.
Grant contrasts this active, inquiry-based learning with the traditional university lecture. He cites research showing that students actually learn less from highly rated, charismatic lecturers than they do from active-learning sessions, even though they feel like they learned more from the lecture. He calls this the dumbstruck effect. The sage on the stage mesmerizes us with a tidy, polished narrative, and we mistake the pleasure of being told a story for the work of understanding it.
Complexifying is deeply appealing as a concept. Yet it assumes a marketplace of ideas populated by good-faith actors. Injecting nuance into a debate works when both sides share a fundamental commitment to reality. When bad-faith actors weaponize complexity to stall necessary action — as the fossil fuel industry did for decades by funding skeptical research to manufacture doubt — complexifying the narrative can actually serve the status quo. Doubt is a vital scientific tool. It is also a highly effective corporate weapon, and Grant doesn’t spend enough time with that uncomfortable overlap.
Scorecards and Safety Nets
A performance culture is obsessed with outcomes. It rewards success and punishes failure. This seems logical, but it creates a fatal vulnerability. It destroys psychological safety. If failure is punished, people will hide their mistakes, silence their doubts, and conform to the hierarchy.
Grant uses the tragedies of NASA’s space shuttle program to illustrate the cost. Before both the Challenger and Columbia disasters, engineers noticed anomalies. O-ring damage and foam shedding. Because previous flights with similar anomalies had returned safely, the deviance was normalized. The culture demanded airtight proof that a launch was unsafe before they would halt it. The burden of proof was on the skeptics, and the skeptics lacked the psychological safety to halt the machine.
To build a learning culture, organizations must separate process from outcome.
You can have a good outcome generated by a terrible process. We usually call this luck. If a bank executive approves a highly risky loan without doing due diligence, and the borrower happens to strike it rich and pay it back, the outcome is positive. In a performance culture, the executive gets a bonus. In a learning culture, the executive gets reprimanded, because the decision-making process was shallow and dangerous.
You can have a bad outcome generated by a rigorous, deep process. We call this a smart experiment. If a medical research team follows all protocols, tests a new drug, and the drug fails clinical trials, the outcome is negative. The process was flawless. The organization learned something valuable.
Grant advocates for the implementation of a rethinking scorecard to evaluate decisions based on the rigor of the process, not just the final result. He praises Amazon’s practice of requiring six-page memos that force authors to document the alternative approaches they considered and rejected.
This structural approach to rethinking is where the book finds its firmest footing. You cannot rely on individuals to constantly maintain a Scientist mindset. The cognitive load is too high. You can design organizational systems that force the Scientist mindset into the workflow — a designated devil’s advocate, a review process that explicitly scores the quality of the assumptions made. The organization can rethink even when the individuals within it are tired, biased, or stubborn.
The concept of psychological safety, pioneered by Amy Edmondson, has become a corporate buzzword. Grant handles it well, noting that safety without accountability breeds complacency. A critique often leveled at the psychological safety discourse is its structural blindness. It treats fear in the workplace as an interpersonal failing rather than a rational response to material power dynamics. You cannot out-mindset a bad organizational structure. If a CEO holds absolute power over livelihoods and has a history of firing dissenters, telling middle managers to foster psychological safety is akin to putting a bandage on a bullet wound. True rethinking at the organizational level requires a redistribution of power, a reality that behavioral science often sidesteps in favor of behavioral nudges.
The Foreclosed Future
There is a moment late in the text where Grant discusses the concept of identity foreclosure. This is the tendency to settle prematurely on a sense of self without exploring alternatives. We ask children what they want to be when they grow up, forcing them to lock onto a career identity before they even understand what work is. We march down paths of escalating commitment, sinking years into medical degrees or corporate ladders because we refuse to rethink a plan we made when we were twenty.
This is the quiet tragedy that Think Again is trying to prevent.
If you are looking for a deep, philosophical exploration of rationality, you might be better served by Julia Galef’s The Scout Mindset, which covers similar territory but focuses more intensely on the internal, solitary mechanics of truth-seeking. Grant’s work is distinctly outward-facing. It is built for the manager, the teacher, the consultant, and the parent. He provides the scripts, the frameworks, and the empirical backing needed to alter the geometry of a conversation — and he proves, convincingly, that certainty is a symptom of intellectual stagnation rather than a sign of intelligence. The tools he offers are difficult to use, and they will fail you when your emotions run too hot. The alternative is to keep defending a map of a world that no longer exists. That first-instinct fallacy from the Scantron sheet — it turns out we carry it into everything.
If you’d like to read the full book in EPUB or MOBI format, feel free to send me an email—I’d be happy to share a free copy with you. Please reach me at: thenovaleaf@gmail.com