Tali Sharot runs the Affective Brain Lab at University College London, and her second book is an attempt to answer a question that has become uncomfortably practical: why does presenting someone with good evidence so often leave them believing the opposite more firmly than before? Her answer draws on her own experiments in neuroscience and behavioural economics, and it amounts to a fairly complete inversion of how most of us try to persuade.
The core argument
The instinct when someone is wrong is to supply data. Sharot’s case is that this is close to the least effective available option, because the brain does not process incoming information neutrally. Existing beliefs act as a filter that determines what gets absorbed, and evidence that conflicts with a prior is scrutinised, discounted, or reinterpreted as support for the original view. In her own work on belief updating, people revised their expectations readily in response to good news and barely at all in response to bad — an asymmetry that makes optimism durable and correction difficult.
Her constructive move is to stop fighting the other person’s mental machinery and start working with it. Rather than attacking a prior belief head-on, find common ground and build the argument from a premise the listener already accepts. Rather than leading with threat, lead with what someone can gain, since fear reliably produces inaction and avoidance while anticipated reward produces movement. Rather than removing choice, preserve the other person’s sense of agency, because control is itself something people value and will defend at the cost of accuracy.
Around this she assembles seven levers: prior beliefs, emotion, incentives, agency, curiosity, state of mind, and the behaviour of other people. The chapter on curiosity is the most useful — information gaps produce a genuine craving, and the anticipation of resolving one activates reward circuitry, which means a well-posed question does work that a well-argued answer cannot.
Key ideas
- The backfire risk of counter-evidence. Contradicting a strongly held belief with data can entrench it. Persuasion works better from shared premises than from opposing ones.
- Emotion is contagious and transmissible. Physiological states synchronise across an audience, which is why a speech can move a room that an identical written argument would not.
- Reward beats threat. Warnings suppress action; the promise of gain drives it. Public-health and safety messaging routinely gets this backwards.
- Agency is a motivator in itself. People will take a worse outcome to keep control over it, so influence that removes choice tends to provoke resistance.
- Curiosity as a lever. Opening an information gap makes people want the answer, and information sought is retained far better than information delivered.
- State of mind gates everything. The same message lands differently on a stressed brain than a calm one; timing is not a garnish on persuasion but part of it.
Who it’s for
This is a good fit for anyone whose work involves changing minds — teachers, clinicians, managers, policy people — and especially for anyone who has been frustrated that being right did not seem to help. Sharot writes cleanly, uses her own lab’s work rather than recycling the usual popular-science canon, and is more careful than most in the genre about the distance between a finding and a life lesson.
The fair criticisms are real, though. It sits in the social-psychology tradition that the replication crisis has hit hardest, and several supporting studies cited across the field — including work on the backfire effect itself — have proved shakier on re-examination than the confident framing suggests; treat individual results as suggestive rather than established. The book is also short and somewhat thin in places, with chapters that end just as they get interesting, and readers already versed in Kahneman or Cialdini will find some of the terrain familiar. As a corrective to the assumption that facts speak for themselves, it is still one of the sharpest books available.