Evidence status: conditional support Across 25 experiments, merely instructing participants to imagine another person’s perspective did not systematically improve predictions of emotions, preferences or attitudes. Obtaining information from the other person did help. These experiments measured predictive accuracy, not caring, stereotype reduction or willingness to approach; perspective taking has independent support for some of those outcomes. This is a version 0.1 research draft, with evidence reviewed to August 2026.
How Everyday Psychology Takes Shape · Season Two: How Do We Read Other People? · Article 11
During disagreement, people often advise: “Put yourself in her shoes.” The sentence has moral force. It asks us to leave our immediate interests temporarily and acknowledge that another person possesses reasons and experience.
Willingness to consider another person and accurate knowledge of what that person thinks are not the same capacity. Someone can imagine with great sincerity and still answer the question, “How would I feel if I were in that position?” Imagination increases engagement but does not automatically add data about the target.
This distinction does not belittle perspective taking. It lets us specify the goal. Are we trying to reduce hostility and increase care, or predict one person’s preference? Imagination may improve the first kind of outcome. The second generally needs information.
Why 25 experiments produced a counterintuitive result
Tal Eyal, Mary Steffel and Nicholas Epley published 25 experiments in 2018 testing whether instructions to take another person’s perspective improved interpersonal accuracy. Tasks included judging emotion from faces and body postures, distinguishing genuine from fake smiles and lies from truths, and predicting preferences, attitudes and reactions among strangers, acquaintances and spouses.
The researchers compared perspective-taking instructions with control conditions and calculated accuracy from the other person’s actual answers or a task criterion. Across the programme, perspective taking did not systematically improve accuracy. A pooled analysis even produced a small negative effect. Individual studies did not all show significant harm, so the result should not be rewritten as “imagination always makes judgment worse”. The sounder conclusion is that a simple instruction did not reliably improve mind-reading accuracy.
The final experiment added perspective getting. Rather than only adopting a perspective mentally, participants gained information through conversation or the target’s answers. Accuracy improved as relevant information became available. The title of the paper captures the distinction: accurate understanding requires getting perspective, not merely taking it.
The programme is strengthened by varied tasks, relationships and cumulative experiments. It also has limits. Most tasks measured immediate prediction under controlled conditions, while real relationships involve learning over time. A brief “imagine” manipulation cannot represent all perspective training, literary engagement or sustained contact.
How imagination projects the self into another person
To simulate someone else, the easiest material to retrieve is our own preferences, values and emotional reactions. When people are similar, projection can generate a good guess. When the relevant difference lies outside awareness, vivid imagination carries the self further into the other person’s place.
Stereotypes can fill gaps. We may believe that we have entered an individual’s circumstances while drawing on a ready-made story about age, occupation, gender or cultural group. A more detailed simulation can increase confidence. Confidence and accuracy remain separable.
Close relationships do not eliminate the problem. History provides extensive baseline knowledge, which is an advantage. It can also create a fixed claim—“I know exactly what you will think”—whose authority rests on past accuracy while the partner has changed.
There is a normative substitution too. When taking perspective, people can answer how the target should feel rather than how the target actually feels. During conflict, the perceiver’s moral framework enters the simulation. The result becomes a reasonable reaction written for another person.
Perspective taking can still have another value
Andrew Todd and Adam Galinsky’s review of intergroup research described studies in which perspective taking was associated with more favourable intergroup evaluations, less stereotyping, greater helping and more approach-oriented behaviour. These effects have contextual qualifications, but they show that failure to improve accuracy does not imply absence of psychological value.
A person may still guess another’s preference incorrectly while becoming more willing to acknowledge that person’s subjective world and listen. Moral attention can create the conditions for obtaining information. The reverse is equally important: accuracy does not guarantee benevolence. A manipulator can understand a vulnerability and use it.
Perspective taking and perspective getting are therefore not competing techniques. The first may alter attention and attitude; the second contributes person-specific evidence. A productive sequence is often to become willing to move beyond self-centred concern, ask the person, and allow the answer to correct the imagination.
This sequence also avoids a false choice between compassion and evidence. Good intentions do not exempt a belief from correction, and correction does not require indifference.
Asking directly does not deliver absolute truth
Another person may not know a future preference, may edit an answer or may provide a safe response under power. Reports change over time. Perspective getting is valuable not because it reaches an infallible ground truth, but because a previously absent source enters the evidence.
Question design matters. “You are angry because you were ignored, right?” has inserted the answer. “What mattered most to you in that moment?” is more open. For material the person does not wish to disclose, respecting non-response is part of understanding. Being an important source does not abolish privacy.
Actual judgments can combine sources: what the person says, how they act over time, what the setting permits, and what prior patterns show. When statement and behaviour conflict, one need not immediately declare deception. Uncertainty can be retained while the important difference is investigated.
In an institutional setting, perspective getting may also require safe channels. Asking employees what they think while their manager is present is not equivalent to creating conditions in which disagreement is possible.
Perspective getting is not compelled disclosure. A person may decline to describe private experience or answer only what is relevant to the present decision. Respecting that boundary does not defeat understanding. It prevents a questioner from treating “I need to know” as “the other person must reveal everything”. The observer remains responsible for the inference and the action built from it.
When a judgment affects employment, discipline, benefits or reputation, direct inquiry is only one part of the evidence structure. Institutions should record how questions were asked, what material supported the conclusion, where uncertainty remains, and how the person can review and contest the account. The institutional value of perspective getting lies not merely in collecting more words, but in giving a mistaken interpretation a trace and a usable route to correction.
From thinking for someone to allowing correction
Perspective taking often preserves an invisible centre: I construct the other person’s world, and I decide whether the simulation is adequate. Perspective getting changes the relational structure. The person being understood can provide information and say, “No, that is not what happened for me.”
This is an epistemic responsibility. A person is not always the best analyst of every cause of their behaviour, but their account of experience has a standing that external guesses cannot casually cancel. Exposing an interpretation to correction is more reliable than constructing an increasingly complete imaginary other.
The version 0.1 judgment is that putting oneself in another’s shoes can interrupt self-centred attention, support care and prepare listening. It should not be treated as a self-sufficient technique for accurate prediction. If the answer matters, ask, listen, observe and allow the person to change the model. Sometimes the most important advance in understanding is not imagining more vividly, but recognising that imagination is still not enough.
Primary research sources
- Eyal, T., Steffel, M., & Epley, N. (2018), “Perspective Mistaking”.
- Todd, A. R., & Galinsky, A. D. (2014), “Perspective-Taking as a Strategy for Improving Intergroup Relations”.
- Galinsky, A. D., & Moskowitz, G. B. (2000), “Perspective-Taking: Decreasing Stereotype Expression, Stereotype Accessibility, and In-Group Favoritism”.
- Epley, N., Keysar, B., Van Boven, L., & Gilovich, T. (2004), “Perspective Taking as Egocentric Anchoring and Adjustment”.
Series navigation: How Everyday Psychology Takes Shape — Season Two overview
Discover more from Geoffrey Chen
Subscribe to get the latest posts sent to your email.