.png)
Decoding Your Executive: What the Research Says About Reading the Person You Support.
Updated: Sep 28
The door opens, your executive walks past your desk without a word, and by the time they reach their office you have already rearranged the afternoon. The budget question can wait until tomorrow. The difficult email can sit in drafts. You know that walk. Then, at four o'clock, they wander back out cheerfully to ask why nobody has raised the budget, and it turns out the silence meant nothing more than a speech they were rehearsing in their head.
Every Executive Assistant and Chief of Staff carries a working model of the person they support: what they like, what rattles them, how they decide, what that particular walk means. Most of the time the model is good, and it is one of the most valuable things you bring to the job. I wrote about decoding your executive in 2025, pairing the instinct you build on the job with some of psychology's frameworks, and this is the fuller version of that idea, built on what the research says about how accurately people read one another, what makes the reading better, and where our confidence runs ahead of it.
What you will get out of this
How accurate people really are at reading another person's thoughts and feelings, and why the people who feel most skilled at it are not reliably the ones who are
The four things a good read needs, and the surprising finding about gut feel from studies of executive-level professionals
What the Big Five traits tell you about executives, which parts of personality you can see better than they can, and which parts only they can see
What the research says about four-letter personality types, and why people are better described as dimensions than as boxes
The if-then map: the way of describing a person that predicts what they will do next far better than a label
Why first impressions last, when they are right, and how they quietly make themselves come true
Why what someone says tells you more than how they look, and why asking beats guessing
How to build a picture of your executive's preferences, stress signals and decision style that you can keep updating without turning it into a stereotype
The reading check: a tool that takes one thing you believe about your executive and tells you how solid it is
The Executive File Pages: three pages for building and keeping an accurate, updatable picture of the person you support
How well we really read each other
Psychologists call the skill at the heart of your job empathic accuracy, and they measure it with a method that is simple and slightly brutal. Two people have a conversation, which is filmed. Afterwards each of them watches the tape alone and stops it wherever they remember having a particular thought or feeling, and writes it down. Then each person watches the tape again and, at the other person's stopping points, writes down what they think the other person was thinking or feeling at that moment. The two lists are compared.
The first large study using this method, published in 1990, filmed 38 pairs of strangers and found that accuracy could be measured reliably, and that it depended a great deal on the interaction itself rather than on either person alone. It also found something that should make all of us slightly humble. People's own ratings of how empathic they were, and how good they were at reading others, did not predict how accurate they turned out to be.
That finding has only grown with time. A 2019 meta-analysis of 85 studies and more than 14,000 people found that self-reported ability to understand others accounted for about one per cent of the variation in how people actually performed on tasks that measured it. A preregistered study of 700 people in 2024 put it even more starkly: people's view of their own mind-reading ability had almost no relationship with their measured ability, a correlation of .05, and was related instead to how well they thought they had done. A study of relationship partners and roommates found that the longer and closer a relationship was, the more confident people became in their impressions of the other person, without becoming reliably more accurate about them. Knowing someone well made the picture richer, and the richness fed the confidence.
So the first finding is sobering: how sure we feel about someone and how right we are about them can come apart without our noticing.
The second finding is more encouraging, and it is the one to hold on to. Knowing someone does make you better at reading them, in a specific way. A study of male friends and male strangers found friends were more accurate at inferring each other's thoughts and feelings, and the reason was not that they talked more or were more alike. It was that friends knew each other's lives, so they could follow a thought to a place or a time the stranger could not see. A study of dating couples, their friends and strangers found the same gradient: accuracy rose with acquaintance, and better accuracy went with more prior disclosure about the problems being discussed. And a pair of studies on what it means to know someone found that people in close relationships are particularly accurate about the things that matter for the relationship itself, which the researchers called pragmatic accuracy, and that this kind of accuracy went with harmony.
That is your advantage, and it is a real one. You are unlikely to be a better all-purpose mind reader than anybody else. You are likely to be very good at reading this one person on the things that matter to working together, provided you keep the knowledge current.
The feeling of knowing someone and the fact of knowing them are two different measurements, and they drift apart more easily than any of us would like.
There is one more twist worth knowing, because it runs directly against intuition. Closeness can make communication worse. A 2011 study named it the closeness-communication bias, finding more egocentrism in communication between friends than between strangers, and a 2021 follow-up found that shared experience and group membership fed the same overconfidence. The better you know your executive, the more likely both of you are to assume the other has understood.
Try this. Pick three things you would confidently say about your executive: a preference, a pressure point and a habit. For each, write down the last time you checked it against something they actually said or did. If the answer is more than a few months ago, or never, it has become a belief rather than a reading, and that is where to start.
The research behind this section
Ickes and colleagues (1990) developed the empathic accuracy procedure in unstructured interactions of 38 mixed-sex dyads, found that content accuracy could be measured reliably and was to a large extent an emergent product of the interaction, and found that self-report measures of empathic skill and empathic accuracy did not predict it. Murphy and Lilienfeld (2019) meta-analysed 85 studies with a total N of 14,327 and found that self-reported cognitive empathy accounted for approximately 1% of the variance in behavioural measures of cognitive empathy. Wendt and colleagues (2024) ran a preregistered multimethod study with 700 participants and found a negligible latent correlation of .05 between mindreading self-reports and mindreading tasks, a large correlation of .85 between tasks and general cognitive ability, and an association of .29 between self-reports and perceived task performance. Ames and Kammrath (2004) reported that narcissism, rather than actual competence, predicted people's estimates of their own mind-reading ability. He and Côté (2023) studied 1,126 students and 2,205 informants and found that measured empathic accuracy, but not self-views of it, related positively to relationship satisfaction, while self-views, but not measured accuracy, related to life satisfaction. Swann and Gill (1997) found in dating partners and college roommates that relationship length and involvement increased the confidence and richness of impressions without consistently increasing their accuracy. Stinson and Ickes (1992) found that male friends were more accurate than male strangers in inferring each other's thoughts and feelings, and that the difference was attributable to knowledge of each other's lives rather than to more interaction or greater similarity. Thomas and Fletcher (2003) found across dating couples, friends and strangers that mind-reading accuracy rose with acquaintanceship, varied with perceivers' ability, and went with more prior disclosure about the problems discussed. Gill and Swann (2004) found that people achieve substantial pragmatic accuracy, meaning accuracy that serves relationship-specific goals, and that it was uniquely associated with relationship harmony. Savitsky and colleagues (2011) reported increased egocentrism in communication among friends compared with strangers, and Van Der Wege and colleagues (2021) found that familiarity, group membership and shared experience fed overconfidence in the same closeness-communication bias. Kenny and Albright (1987) set out the social relations approach to accuracy and distinguished individual accuracy, how a person behaves in general, from dyadic accuracy, how they uniquely behave with you.
What makes a better reader
If confidence is a poor guide, what actually makes someone read another person well? The most useful framework comes from a 1995 paper on the accuracy of personality judgement, which set out what it called the realistic accuracy model. For a read to be accurate, four things have to happen in order. The person has to do something that is relevant to the trait or state in question. That behaviour has to be available to you, which means you have to be there, or be told. You have to detect it. And you have to use it correctly. A failure at any one step breaks the whole chain.
That model explains a great deal of what the later research found. Studies that varied how much and what kind of information people had about a stranger found that both the quantity and the quality of information raised accuracy, which is the availability step. A study of unacquainted groups found that good judges tended to be socially skilled, agreeable and well adjusted, and seemed to draw revealing behaviour out of the people around them. A 2019 set of four large studies found that some people really are better judges of others, and that the difference showed most clearly when the person being judged was easy to read. Some people are easier to read than others: a review of what researchers call judgeability found the easier ones tended to be well adjusted and socially skilled, and a study of emotional reading found that perceivers' empathy only helped when the person they were reading was expressive.
Which is worth knowing if you support someone who gives very little away. It may be harder to read them than it was to read your last executive, and that says more about the two of them than it does about you.
Then comes motivation, which turns out to matter more than most people expect. A 2000 meta-analysis of 15 empathic accuracy studies found that the familiar gap in which women outperform men only appeared when people knew they were being assessed on empathy, or when gender expectations were made obvious, which pointed to motivation rather than ability. A 2001 pair of studies confirmed it: paying people for accuracy improved both men's and women's performance and removed the difference altogether. That result has a complication, and it is fair to give you both halves. A 2013 study reported what its title calls an ironic effect of financial incentive on empathic accuracy, finding that paying people to be accurate could backfire. My reading is that caring about getting it right helps, while pressure to perform can get in the way of the attention that makes reading possible.
The most useful single study for this profession looked at gut feel directly. Across four studies, three of them using executive-level professionals, researchers found that most people believe empathic accuracy comes from intuition. The evidence ran the other way. People who tended to rely on intuitive thinking were less accurate at reading others' emotions, face to face and from photographs, and when the researchers induced a more systematic way of thinking, accuracy improved, which showed the effect was causal.
None of this argues against instinct, which in your case is built from hundreds of observations and is often right. It argues for giving the instinct a second look before you act on it: what you saw, what else it could mean, and whether you have seen it before in the same kind of situation.
The last ingredient is practice, and here the evidence is encouraging. A 2012 meta-analysis of training to improve person perception found that training worked, that the type of training mattered more than its length, and that practice and feedback worked better than instruction alone, with a combination working best. The feedback part matters most in your job, because it is the part most people never get. You make a read, you act on it, and unless you check, you never find out whether you were right.
And the reason all of this is worth the effort: accurate reading pays. A study of 142 employees, their peers and their supervisors found that the ability to recognise emotions was linked to annual income, through political skill and the ability to smooth working relationships. In two studies of salespeople, those who were better at reading nonverbal emotion earned bigger salary increases and sold more cars, even though they did not rate themselves as more successful. A study of 145 managers found emotion recognition predicted transformational leadership behaviour. A review of what interpersonal accuracy actually does in conversations concluded that there is no single behaviour associated with it; what accurate readers do is adapt, changing what they do to fit what the other person needs.
Try this. Keep a read log for two weeks. When you act on a read of your executive, such as holding back a question because of their mood, or bringing forward a decision because you sense they are ready, write one line: what you saw, what you concluded, what you did. At the end of each week, check three of them. Were you right? The log is the feedback loop the training research says makes the difference.
The research behind this section
Funder (1995) introduced the realistic accuracy model, describing accurate personality judgement as a function of the relevance, availability, detection and utilisation of behavioural cues. Letzring, Wells and Funder (2006) manipulated information quantity and quality in triads of unacquainted students and found both were positively related to realistic accuracy. Letzring (2008) found that accuracy was related to the judge's social skill, agreeableness and adjustment, and that observers were more accurate when more good judges took part in the interaction. Rogers and Biesanz (2019) found across four large samples that good judges were better able to detect and use valid cues, but that this was strongly evident only when judging good targets. Human and Biesanz (2013) reviewed judgeability and identified psychological adjustment, social status and socialisation as characteristics of people who are easier to read accurately. Zaki, Bolger and Ochsner (2008) found that perceivers' trait affective empathy predicted empathic accuracy only for expressive targets. Ickes, Gesn and Graham (2000) meta-analysed 15 empathic accuracy studies and found that gender differences appeared only where empathy was made salient, supporting a motivational rather than ability explanation. Klein and Hodges (2001) found that payment for accuracy improved the performance of both men and women and eliminated the gender difference. Ma-Kellams and Blascovich (2013) reported an ironic effect of financial incentive on empathic accuracy. Ma-Kellams and Lerner (2016) ran four studies, three with executive-level professionals, and found that contrary to lay beliefs, a tendency towards intuitive thinking went with lower empathic accuracy, and that experimentally inducing systematic thinking improved it. Blanch-Hartigan, Andrzejewski and Hill (2012) meta-analysed training to improve person perception accuracy and found that training significantly increased accuracy, that approach mattered more than length, and that practice and feedback were more effective than instruction alone, with combined approaches most effective. Momm and colleagues (2015) studied 142 employee, peer and supervisor triads and found emotion recognition ability indirectly related to annual income through political skill and interpersonal facilitation. Byron, Terranova and Nowicki (2007) found that building-supply salespeople better at nonverbal emotion recognition earned higher salary increases and car salespeople better at it sold more cars per month, while neither group rated themselves as more successful. Rubin, Munz and Bommer (2005) studied 145 managers and found that emotion recognition, positive affectivity and agreeableness predicted transformational leadership behaviour. Schmid Mast and Hall (2018) reviewed the behavioural correlates of interpersonal accuracy and concluded that it fosters behavioural adaptability rather than any one specific behaviour.
The five traits, and the part of them you can see
When psychologists describe personality, they mostly use five broad dimensions: openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism, which is often described from the other end as emotional stability. They are dimensions rather than categories, which means everybody sits somewhere along each one, and they have a large body of evidence behind them.
For executives, the most useful study is a 2002 meta-analysis of 222 correlations from 73 samples linking the five traits to leadership. Extraversion had the strongest relationship at .31, followed by conscientiousness at .28, openness at .24 and emotional stability at .24, with agreeableness barely registering at .08. Together the five had a multiple correlation of .48 with leadership. The traits help explain who rises into senior roles, which is a reminder that the average executive is a little more outgoing, organised, open and steady than most, with a wide spread around each average.
Personality at the top shapes more than the executive's own day. A study of 17 chief executives found that their personality related to how their top teams worked together, and that the team dynamics related to organisational performance. A study of 195 firms found chief executives' personality traits shaped how quickly their firms adapted to change, with each trait facet either helping or hindering. A 2019 study developed a way of measuring chief executives' Big Five traits from their language, applied it to more than 3,000 chief executives, and found their traits had a meaningful effect on strategic change, with the effect depending on how the firm had recently performed. All of this sits inside a long tradition, the upper echelons view, which holds that organisations partly reflect the people at the top.
A few pieces of evidence tell you how much weight to put on a trait-based picture. The first is stability over time. A meta-analysis of 152 longitudinal studies found that the consistency of personality traits rises through life, from .31 in childhood to .54 in the college years and .64 at age 30, reaching a plateau around .74 between 50 and 70. Your executive's traits are fairly stable, though never fixed, and a review of life events found some evidence that major events in love and work can shift them. The second is whether the findings hold up when repeated. A 2019 project repeated 78 published links between traits and life outcomes with large samples, and 87% of them replicated, typically at about 77% of their original strength. The Big Five map is real, and a little less dramatic than the first studies suggested.
The third piece of evidence is the one I find most useful for anyone in a support role. It turns out that different people are best placed to see different parts of a personality. A 2010 study had people rate themselves and be rated by friends and strangers, and then tested everyone against behavioural measures. People were the best judges of their own neuroticism-related traits, the inner, less visible ones such as anxiety. Friends were the best judges of intellect-related traits, the more evaluative ones, which people tend to be less objective about in themselves. Extraversion was judged equally well by everybody.
The same pattern holds at scale. A set of meta-analyses covering more than 44,000 people rated by others found that frequent contact improved accuracy, but that it took real closeness to improve it substantially, especially for low-visibility traits such as emotional stability. And a meta-analysis of observer ratings found that other people's ratings of someone's personality predicted their job performance better than their own self-ratings did.
You probably see how your executive comes across more accurately than they do. They almost certainly know what is going on inside them better than you do.
Put that together and you have a practical rule about stress signals, which is where most EAs and Chiefs of Staff want the model to be sharpest. You are well placed to see the outward pattern: the shorter emails, the cancelled one-to-ones, the way the diary gets defended. You are poorly placed to see the inner state that drives it, because anxiety is precisely the trait that is least visible from outside. The accurate way to learn what stress feels like for them, and what helps, is to ask when things are calm.
Try this. On one page, describe your executive on the five dimensions in your own words, using behaviour you have seen rather than adjectives. Then mark each line with who is best placed to know it. The outward ones you can hold with some confidence. The inner ones, especially anything about stress, go on a list of things to ask about in a quiet moment.
The research behind this section
Judge and colleagues (2002) meta-analysed 222 correlations from 73 samples and found correlations with leadership of −.24 for neuroticism, .31 for extraversion, .24 for openness, .08 for agreeableness and .28 for conscientiousness, with a multiple correlation of .48, and extraversion the most consistent correlate across settings and criteria. Barrick and Mount (1991) found that conscientiousness showed consistent relations with job performance across occupational groups, and that extraversion predicted performance for managers and sales. Hogan and Kaiser (2005) reviewed the evidence and argued that personality predicts leadership and that leadership is about the performance of teams and organisations. Peterson and colleagues (2003) studied 17 chief executives and found relationships between chief executive personality and top management team dynamics, and between those dynamics and organisational performance. Nadkarni and Herrmann (2010) studied 195 firms and found that chief executive personality drove strategic flexibility, with each facet enhancing or inhibiting it. Harrison and colleagues (2019) developed and validated a linguistic measure of chief executives' Big Five traits, applied it to more than 3,000 chief executives of S&P 1500 firms, and found the traits had a meaningful impact on strategic change that differed with recent firm performance. Kaplan, Klebanov and Sorensen (2012) found candidates for chief executive roles varied on general ability and on a dimension contrasting interpersonal with execution skills, and that subsequent performance was related to general ability and execution skills. Hambrick and Mason (1984) proposed the upper echelons perspective, that organisational outcomes are partially predicted by managerial characteristics. Roberts and DelVecchio (2000) meta-analysed 152 longitudinal studies and found rank-order trait consistency rising from .31 in childhood to .54 in the college years, .64 at age 30 and a plateau of about .74 between ages 50 and 70. Bleidorn, Hopwood and Lucas (2018) reviewed prospective research on nine major life events and found some, still preliminary, evidence that life events can change personality traits. Soto (2019) ran preregistered, high-powered replications of 78 trait-outcome associations and found 87% replicated, with effects typically 77% as strong as the originals. Vazire (2010) tested the self-other knowledge asymmetry model with 165 participants rated by friends and strangers, and found the self the best judge of neuroticism-related traits, friends the best judges of intellect-related traits, and all perspectives equally good at extraversion-related traits. Connelly and Ones (2010) meta-analysed ratings of 44,178 targets across 263 samples and found that interaction frequency improved accuracy but that interpersonal intimacy was needed for substantial gains, especially for low-visibility traits such as emotional stability. Oh, Wang and Mount (2011) found that observer ratings of the five traits predicted overall job performance better than self-reports, and added to self-reports while the reverse was not true. Kenny and West (2010) meta-analysed 24 studies of agreement and similarity in self and other perception.
Why four letters are not a person
Most people who have worked in an office for long enough have been handed four letters, or a colour, or a single capital from a four-box model, and told that it describes them. The appeal of these labels is obvious. A type is memorable, it gives a team a shared vocabulary, and it feels like a shortcut to knowing someone. It is worth knowing what the evidence says about that shortcut.
The most widely used type indicator has been reviewed repeatedly. A 1989 study of 468 adults found no support for the idea that it measures truly two-sided preferences or distinct types; it measures four relatively independent dimensions, and those dimensions line up with four of the five traits in the Big Five. Reviews in 1993, 1995 and 2005 concluded that the evidence did not support the claims made for the test's usefulness, and cautioned against inferences from the four-letter formula.
There is a fair counterweight, and I want to give it properly. The scores themselves are reasonably reliable. A reliability study across many administrations found strong internal consistency and test-retest reliability, with some variation, and a 2025 synthesis of 193 studies found internal consistency between .85 and .92 and robust overlap with six other personality instruments. The scales measure something real. The same 2025 synthesis also noted that across 25 years of literature there were no studies of the test's structural validity or of test-retest reliability for the current form. The dimensions underneath are sound; the step from dimensions to boxes is where the trouble lies.
The deeper question is whether people come in types at all. Two large reviews of taxometric research, the statistical method designed to tell categories from continua, give a very clear answer. A 2012 review of 177 articles found that 38.9% of findings looked like categories, but that these were much rarer in the more recent and methodologically stronger studies, and that after controlling for method the true figure was about 14%. Normal personality showed little persuasive evidence of types. A 2020 meta-analysis found findings supporting dimensions outnumbered those supporting categories by five to one, and that no content area showed overall support for types.
Why do types feel so accurate when you read your own profile? Part of the answer is a very old finding. In a 1949 classroom demonstration, students rated a personality sketch as a strikingly accurate description of themselves, and every student had been given the same sketch. A review of the research that followed found that people readily accept vague, general and favourable statements as describing their unique personality, especially when the statements seem relevant and flattering. It is called the Barnum effect, and a well-written type description is almost perfectly designed to trigger it.
The four-box behavioural models that describe people as dominant, influential, steady or conscientious rest on the same logic, and I could find much less independent peer-reviewed testing of them. Everything above about dimensions and boxes applies.
The conversation a type indicator starts can still be useful, because it tells you something about how your executive likes to see themselves. The mistake is treating the label as a forecast. Two people with the same four letters can be far apart on every underlying dimension, because a box draws a line through the middle of a continuum and puts everyone near the line in whichever side they happened to land on.
Try this. If you have a type or colour for your executive, translate it back into behaviour. For each letter or colour, write one thing you have actually seen them do that fits it, and one thing you have seen them do that does not. The second column is where the useful information is, because it tells you where the label stops working.
The research behind this section
McCrae and Costa (1989) evaluated the Myers-Briggs Type Indicator with 267 men and 201 women using self-reports and peer ratings, found no support for the view that it measures dichotomous preferences or qualitatively distinct types, found instead that it measures four relatively independent dimensions, and found that the four indices measured aspects of four of the five major dimensions of personality. Pittenger (1993) evaluated the indicator using a unified view of test validity and concluded there was insufficient evidence to support its tenets and claims of utility. Boyle (1995) reviewed its psychometric limitations and recommended against routine use. Pittenger (2005) set out several psychometric limitations that warrant considerable caution when making inferences from its four-letter type formula. Capraro and Capraro (2002) conducted a reliability generalisation study and found generally strong internal consistency and test-retest reliability estimates, with variation across administrations. Erford and colleagues (2025) synthesised 193 studies of Form M from 1999 to 2024 and found internal consistency of .845 to .921 and robust convergent evidence with six instruments, and found that structural validity and test-retest studies were absent from the 25-year literature. Haslam, Holland and Kuppens (2012) reviewed 177 taxometric articles and 311 findings, found 38.9% taxonic but far fewer in recent and stronger studies, estimated the true prevalence of taxonic findings at 14%, and found little persuasive evidence of taxa in normal personality. Haslam and colleagues (2020) meta-analysed 317 findings from 183 articles and found dimensional findings outnumbered taxonic ones five to one, with no content domain showing aggregate support for the taxonic model. Forer (1949) demonstrated in a classroom that people accept general personality descriptions as accurate descriptions of themselves, which he called the fallacy of personal validation. Dickson and Kelly (1985) reviewed research on the Barnum effect and concluded that acceptance of vague, general statements depends on their relevance and favourability and to some extent on the type of assessment.
People in situations: the if-then map
If types do not predict well, what does? The most useful answer I know comes from a line of research that started with a puzzle. Personality is fairly stable, and yet people behave differently from one situation to the next. A 1995 theory resolved it by suggesting that the stable thing about a person is not an average level of behaviour but a pattern: if this kind of situation, then this, but if that kind of situation, then that. The researchers called these behavioural signatures.
A study of children at a summer camp, observed across many situations, found exactly that: the way each child's behaviour varied across situations was itself stable and distinctive, a profile rather than a level. A 2001 study that sampled adults' behaviour several times a day found that people's behaviour ranged widely around their average on every trait, and that the amount of variation in extraversion reflected how strongly each person reacted to cues in the situation. A trait, in that view, is better thought of as a spread of states than as a fixed setting.
We already think this way about people we know well: a 2005 set of studies found that people link common trait words such as friendly and shy to characteristic if-then patterns, and use a person's stable pattern to infer their motives. It is in the office, with labels to hand, that we sometimes forget.
Some years ago I worked alongside an EA whose executive was, by every measure, a creature of structure, apart from one thing: urgent requests arrived every Friday afternoon, reliably wrecking her carefully organised week. She looked back over a couple of months of them and found the pattern. Friday was the first time all week her executive had room to think, and the requests were the week's thinking arriving all at once. She started sending a short note on Wednesday mornings asking whether anything might be needed by Friday, and most of the Friday panic went away. That is an if-then signature, found by paying attention, and it was far more useful than any label.
The research on work adds two refinements. A trait-activation model proposed that traits show up at work when the situation calls for them, and a review of situational strength found that strong situations, the ones with clear rules, consequences and expectations, restrict how much personality shows at all. In a board meeting with a set agenda, you are seeing the situation more than the person. In the unstructured hour after a difficult call, you are seeing the person more than the situation.
A label tells you what your executive is like on average. An if-then map tells you what they will do next Tuesday at four.
There is one bias to watch for, and it is among the best documented in psychology. The correspondence bias is the tendency to put behaviour down to a person's enduring character when the situation could entirely explain it. A review of the evidence identified four routes into it, among them not being aware of the situational pressures at all and failing to correct for them when you are. From your seat, it looks like this: the sharp email at 11pm gets filed under "impatient", when the situation was a board paper due at midnight and a sick child upstairs.
Try this. Take one thing you think of as a trait of your executive, such as "hates detail" or "decides fast", and rewrite it as two if-then lines, one where it holds and one where it does not. "If it is an operational decision and the numbers are clear, they decide in the meeting. If it touches people or reputation, they want a night on it." Two lines like that will serve you better than any single word.
The research behind this section
Mischel and Shoda (1995) proposed a cognitive-affective system theory of personality that accounts for stable, predictable patterns of variability across situations, such as if A then she X but if B then she Y, as behavioural signatures of personality. Shoda, Mischel and Wright (1994) found in observations across situations that the intraindividual organisation of behaviour variation was enduring and discriminatively patterned, visible as distinctive profiles of situation-behaviour relationships. Fleeson (2001) found that individuals' behaviour varied widely around their trait levels, described traits as density distributions of states, and found that within-person variability in extraversion reflected reactivity to situational cues. Kammrath, Mendoza-Denton and Mischel (2005) found that perceivers associate common trait terms with characteristic if-then signatures and use a target's stable signature to infer motives and traits. Tett and Burnett (2003) proposed a trait-based interactionist model of job performance distinguishing job demands, distracters, constraints, releasers and facilitators as features relevant to trait expression. Meyer, Dalal and Hermida (2010) reviewed situational strength, organising it into four facets: constraints, consequences, how clear the expectations are, and consistency, and set out the idea that strong situations restrict the expression of individual differences. Gilbert and Malone (1995) reviewed the correspondence bias, the tendency to infer enduring dispositions from behaviour that can be explained by situations, and described four mechanisms producing it: lack of awareness, unrealistic expectations, inflated categorisations and incomplete corrections.
First impressions, and why they stick
First impressions deserve more respect than they usually get, and more caution too.
Start with the respect they deserve. A study of 600 people rated from fifteen short videotaped episodes found that personality judgements from thin slices of behaviour were significantly associated with the views of people who knew them well. The association strengthened as more episodes were added, but gains were low beyond about six. A study of new classmates found that forming an accurate first impression, independent of simply liking the person, predicted more interaction over the semester and greater liking by the end of it. Accuracy at the start was good for the relationship.
Now for the caution, which is about persistence. First impressions persist, whether or not they are right. In a 2017 study, people judged a stranger from a photograph and, at least a month later, met the same person face to face. Their liking from the photograph predicted their liking after the conversation, and part of the reason was that people with a favourable first impression behaved more warmly, which drew more warmth back. A review of social judgements from faces found strong agreement about what faces seem to say, surprisingly fast judgements, and real consequences, and concluded that the accuracy of these judgements has been greatly overstated.
The mechanism behind the persistence has been known since 1978. In four studies, people given a hypothesis about someone else, that the person was an extravert or an introvert, chose questions that would confirm it, and the questions shaped the conversation so that the other person really did behave in a confirming way. A classic study of the halo effect found that people who saw a lecturer being warm rated his appearance, mannerisms and accent as appealing, while those who saw him being cold rated the same things as irritating. They were unaware of the influence and believed it ran the other way.
That is the real risk with a new executive: the first read shapes your behaviour, your behaviour shapes theirs, and a few weeks later you have evidence for a conclusion you drew on the first morning.
The way out is time, used well. A 2007 study found that as people got to know each other, their judgements moved away from generic expectations and towards the distinctive features of that particular person, while a review of 32 studies found that agreement between different people's judgements did not rise with acquaintance, so what improves is the individual picture.
There is a real disagreement here, and it matters for the no-stereotyping part of your model. A review of stereotype accuracy argued that people's beliefs about groups are often more accurate than they have been given credit for, and that using group information is not always irrational. Other research points the other way on what matters most for you. A 1987 set of experiments found that when people's outcomes depended on someone else, they formed impressions from that person's individual attributes rather than from a category, and paid more attention to the individual information. And a 1993 analysis of power and stereotyping found that the people with power tend to pay less attention to those below them and so rely more on stereotypes. My reading is that whatever the average accuracy of a stereotype, your job depends on the individual, and dependence on someone is exactly the condition in which the research finds people look past the category.
The research on power has a disagreement of its own, and it concerns your executive. In one set of experiments, feeling powerful reduced perspective taking and accuracy at identifying emotional expressions; in another, across four studies, it increased interpersonal sensitivity, particularly with an empathic leadership style. Power seems to amplify whichever orientation a leader brings.
Try this. Write down your first impression of your executive, or of anyone new you have started working with, in three lines. Date it and put it away. In three months, read it again and mark what held, what changed and what you might have made come true. It is one of the most useful ten-minute exercises I know.
The research behind this section
Borkenau and colleagues (2004) collected self-reports, peer reports and ratings from 15 videotaped episodes for 600 participants, found an average cross-situational consistency of trait impressions of .43, found that personality inferences from thin slices were significantly associated with reports by knowledgeable informants, and found that gains in prediction were low beyond six episodes. Human and colleagues (2013) found that more accurate first impressions of new classmates, assessed as distinctive self-other agreement, predicted greater interaction throughout the semester and greater liking and interest by its end, independently of initial liking. Gunaydin, Selcuk and Zayas (2017) found that liking judgements from a photograph predicted liking after a face-to-face interaction at least a month later, partly through behavioural confirmation, and that photograph-based personality judgements predicted interaction-based ones to a lesser extent. Todorov and colleagues (2015) reviewed social attributions from faces and argued that their diagnostic validity has been greatly overstated. Snyder and Swann (1978) found across four investigations that people testing hypotheses about others preferentially searched for confirming evidence and channelled interactions so that targets provided behavioural confirmation. Nisbett and Wilson (1977) found that a lecturer's warm or cold manner altered ratings of his appearance, mannerisms and accent, and that subjects were unaware of the influence and believed it ran in the opposite direction. Biesanz, West and Millevoi (2007) found that as acquaintance increased, self-other agreement and differential accuracy increased while stereotype accuracy decreased. Kenny and colleagues (1994) reviewed 32 studies, found consensus correlations ranging from zero to about .3, higher for extraversion, and found no evidence that consensus increases with acquaintance. Cronbach (1955) analysed how scores on understanding of others are affected by assumed similarity and other components. Jussim, Crawford and Rubinstein (2015) argued that stereotype accuracy is one of the largest and most replicable findings in social psychology and that stereotypic information can sometimes be used rationally in person perception. Neuberg and Fiske (1987) found in three experiments that short-term, task-oriented outcome dependency led people to use individuating processes and to attend more to a person's particular attributes, even in conditions that usually produce category-based impressions. Fiske (1993) set out a theory of the mutually reinforcing interaction between power and stereotyping, mediated by attention. Galinsky and colleagues (2006) found that high power reduced visual perspective taking, led people to anchor on their own privileged knowledge, and reduced accuracy in identifying others' emotional expressions. Schmid Mast, Jonas and Hall (2009) found across four studies that high power resulted in more interpersonal sensitivity than low power, partly through feeling respected and proud, and that high-power individuals adopting an empathic leadership style were more interpersonally sensitive.
Ask first, then read
If there is a single practical lesson in this literature, it is this: what a person says tells you more about what they are thinking than how they look while saying it. I covered the perspective-taking side of this when I wrote about staying ahead of your executive; here I want to look at the channels themselves.
A 2007 study presented the same recordings in different forms: full video, audio only, a transcript, or silent video. Verbal information contributed the most to accuracy, followed by the voice. The face and body contributed the least, though still something. An earlier study of therapy sessions found the same order, and added a warning about context. Knowing the background of a conversation helped people follow it, and sometimes led them to answer from their own assumptions about the story rather than from what the person was actually saying.
The evidence on faces explains why they are a weaker channel than we expect. A 2021 meta-analysis tested whether the six classic emotions reliably produce the facial expressions they are supposed to. The average co-occurrence was .13 when full expressions were measured, and .23 including partial ones, with a correlation of .30 between how strongly an emotion was felt and how strongly it showed. A meta-analysis of emotion recognition across cultures found that emotions were recognised better than chance everywhere, and better still when the person expressing and the person reading shared a culture, an advantage that shrank with more contact between groups. So if you and your executive come from different backgrounds, give yourself more time to learn their face, and lean more on their words.
Then there is the other direction, which catches executives as often as their assistants. People overestimate how much of their inner state others can see, an illusion of transparency found across three sets of experiments, and a review of the related research found that closeness and self-awareness make it worse. Five experiments on email found that people believed they conveyed tone far better than they did, because they could hear their own intent. And a review of how people read what others think of them found that people mostly infer it from their own view of themselves, and are better at knowing how others see them in general than how one specific person sees them.
Put those together and you have the most common misunderstanding in any executive office. Your executive assumes you know they are fine, because they know they are fine. You read the silence as a signal, because silence is all you have. Neither of you has said anything.
A 2025 study of 2,225 people added one more reason to ask: people's own values coloured how they saw the values and behaviour of those close to them, even controlling for what the other person actually valued or did. We tend to see in the people we like what we would do ourselves. A review of egocentric thinking found that people use their own minds as the starting point when inferring others' and rarely adjust far enough.
The good news is that asking is not an admission of failure. Six experiments found that believing another person had taken your perspective increased liking for them, closeness to them, and willingness to help them. And a study of how people read others found that accuracy improved when they matched the level of detail at which the other person was thinking, which in practice is what a well-chosen question does. "How are you feeling about this afternoon?" gets an answer at the level your executive is thinking about it, rather than the level you are guessing at.
A diary study of couples found people read their partner's negative moods more accurately than their positive ones, and a study of couples in conflict found that reading threatening thoughts accurately lowered the reader's own sense of closeness. Your antennae for trouble may be the best-tuned part of your model, and accuracy about the hard things has a cost, which is one more reason to check a reading rather than carry it alone.
Try this. Build one standing question into your regular time with your executive that gets you their perspective directly, at the level they are thinking. "What is taking up most of your head this week?" works well, because it asks about their inner weather in a way that is easy to answer. Then read the week in the light of what they said, rather than the other way round.
The research behind this section
Hall and Schmid Mast (2007) presented empathic accuracy stimuli in full video, audio, transcript and silent video and found that verbal information contributed most to accuracy, followed by vocal nonverbal cues, with visual nonverbal cues contributing least though above zero. Gesn and Ickes (1999) found in simulated psychotherapy sessions that empathic accuracy depended primarily on verbal rather than nonverbal cues, and that knowledge of the cumulative context led perceivers to respond in a schema-driven way that sometimes helped and sometimes impaired accuracy. Durán and Fernández-Dols (2021) meta-analysed studies of the co-occurrence of emotion and predicted facial expression and found average co-occurrence of .13 for full expressions, .23 including partial expressions, and a correlation of .30 between reported emotion intensity and expression intensity. Elfenbein and Ambady (2002) meta-analysed emotion recognition within and across cultures and found universal recognition above chance, an in-group advantage, and a smaller advantage for groups with more exposure to one another. Gilovich, Savitsky and Medvec (1998) found across three sets of studies an illusion of transparency, a tendency to overestimate how far others can discern one's internal states. Cameron and Vorauer (2008) reviewed how metaperceptions are egocentrically biased, the factors such as closeness and self-awareness that increase transparency overestimation, and its consequences for relationships and support. Kruger and colleagues (2005) found in five experiments that people overestimated how well they could communicate tone and emotion over email, and that the overconfidence was born of egocentrism. Kenny and DePaulo (1993) reviewed eight social relations studies with 569 participants and found that people determine how others view them largely from their own self-perceptions, and are better at knowing how others generally view them than how specific individuals do. Carlson, Vazire and Furr (2011) found in three studies that people nonetheless have some real meta-insight, making valid distinctions between how they see themselves and how others see them. Ginosar Yaari and colleagues (2025) found in samples totalling 2,225 people that perceivers' personal values related positively to their perception of close others' values and behaviours, controlling for the targets' actual values and behaviours, with some evidence the bias was stronger in better relationships. Todd and Tamir (2024) reviewed the factors that amplify and attenuate egocentric mentalizing, in which people use their own minds as a starting point for inferring others'. Goldstein, Vezich and Shapiro (2014) found across six experiments that perceiving that one's perspective had been taken increased liking for, self-other overlap with, and helping of the perspective taker. Eyal and Epley (2010) found that accurately reading others' minds requires matching the level of construal at which the other person evaluates, and that this was more effective than perspective taking in their context. Howland and Rafaeli (2010) used three weeks of electronic dyadic diaries from cohabiting partners and found greater accuracy for negative than positive target moods. Simpson, Oriña and Ickes (2003) found in married couples that when a partner's thoughts were relationship-threatening, greater empathic accuracy was associated with declines in the perceiver's feelings of closeness, and the reverse when they were not.
A model you can update
So what does an accurate, fair, useful picture of your executive look like? The research points to something closer to a working file than a profile: specific, sourced, dated and expected to change.
It has three parts that matter most in this role.
Preferences, as if-then lines. How they like information, when they like it, and what changes that. The if-then research says to write these as patterns tied to situations, and the realistic accuracy model says to note where each one came from: something they told you, something you have seen many times, or something you have inferred. The last kind is the one to check first.
Stress signals, in two columns. What you can see, such as the outward signs you have learned to recognise, and what they have told you about how pressure feels and what helps. The research on who sees what says you will be better at the first column than they are, and they will be far better at the second.
Decision style, by type of decision. A 1995 study developed a measure of decision-making style with five distinct styles: rational, intuitive, dependent, avoidant and a fifth the researchers found along the way, spontaneous. Most people use more than one, and which one appears often depends on the decision. A related measure of the need for cognitive closure captures how strongly someone wants a firm answer quickly and how uncomfortable ambiguity makes them. Both are more useful as descriptions of what you see across different decisions than as scores to give a person. Some executives decide operational matters in the room and sit on anything to do with people for days, and knowing that is worth more than knowing their average.
Then two habits keep the file honest.
The first habit is updating as things change. Personality is fairly stable in adults, but behaviour varies with situations, and life events can shift even the traits. A new role, a board under pressure, or something happening at home will change what you see. The research on acquaintance found that knowing someone better improves your read of what is distinctive about them, provided the new information keeps coming in.
The second is checking the confident entries hardest. The confidence research is clear that the beliefs you hold most firmly are the ones you are least likely to test, and that familiarity builds confidence faster than it builds accuracy. The research on individual impressions offers the fix: pay attention to the individual rather than the category, which is exactly what depending on someone makes people do when they let it.
An accurate picture of your executive is less like a portrait and more like a weather record: specific, dated, and always being added to.
This file is a working record rather than a dossier or a diagnosis, and that is worth saying plainly. Everything in it should be something you would be comfortable showing your executive, and the best versions of it are built with them. A good way to start that conversation is a simple question: "can I check a few things I think I know about how you like to work?"
Try this. Start the file with five lines: two preferences as if-then patterns, one stress signal you can see, one thing that helps when they are under pressure (only if they have told you), and one note on how they make a particular kind of decision. Put a date and a source on each line. In a month, check the oldest line with them.
The research behind this section
Scott and Bruce (1995) conducted a multistage, four-sample study to develop a measure of rational, avoidant, intuitive and dependent decision-making styles, with internal consistency from .68 to .94, and identified a fifth, spontaneous style in the process. Webster and Kruglanski (1994) developed and validated the Need for Closure Scale as a measure of individual differences in the desire for a firm answer. The remaining studies drawn on here, on if-then signatures, the realistic accuracy model, who sees which traits, stability and change in personality, acquaintance, confidence and individuating impressions, are set out in the earlier sections.
The reading check
Every idea in this post comes down to one moment: you believe something about your executive, and you are about to act on it. Here is a tool for that moment.
Put in one thing you believe about your executive, then answer a few questions about where the belief came from, whether it is about something you can see or something inside them, how many situations you have seen it in, and whether you have ever checked it with them. It will tell you how solid the reading is, what the research suggests could be skewing it, how to firm it up, and it gives you a line for your executive file to copy.
Where to start
If you take one thing from all of this, make it the difference between reading and assuming. Your instinct about your executive is built from hundreds of real observations, and much of it will be right. The research adds only that the feeling of certainty is not a measure of accuracy, that the inner states are the hardest to see from outside, and that the most reliable way to read someone is to ask them and listen carefully to what they say.
After that, there are three moves worth making.
Rewrite your firmest labels as if-then lines, so they describe your executive in situations rather than in general. Build one standing question into your regular time together that gets their perspective directly. And keep a small, dated file of what you know and how you know it, checking the oldest and most confident lines first.
I have put the three pages I would use into a download: an if-then map for turning labels into patterns, a preferences and pressure page that separates what you have seen from what you have been told, and a decision and check-in page for the conversation that keeps the whole picture current.
Reading the person you support well is one of the quietest skills in this profession and one of the most consequential. It is also, the research says, a skill in the full sense: it responds to attention, practice and feedback, and it improves most in the people willing to check it. That is a generous finding for anyone who has ever been told they have a sixth sense for their executive. It means the sixth sense can be trained.
The part of this that shows up in your work sits inside what the Connected Leader Assessment calls Influence and Stakeholder Trust: how well you understand the people you work with, how accurately you anticipate what they need, and whether they trust you to represent them when they are not in the room. It is the dimension that turns knowing your executive into working well with them.
The assessment is free, takes about twelve minutes and gives you a read across all five dimensions. If you want to go further on this one, the Skill Sprints are short, focused programmes on a single dimension, and the Irreplaceable Audit is the place to look at which parts of your understanding of your executive and their world the organisation could not replace tomorrow, which is exactly the knowledge most worth keeping accurate.
Stay savvy.
Meg ✌️
Free tools
The reading check, the interactive tool in this post, free
The Executive File Pages, the download in this post, free
The Connected Leader Assessment, twelve minutes, free, with a read on Influence and Stakeholder Trust, the dimension this post sits in
P.S. Want to Go Deeper on Decoding Your Executive?
If this resonated and you want to go further, here is the evidence the piece is built on. These are the papers that shaped my thinking on how accurately people read one another, and how to build a truer picture of the person you support, and they are well worth your time.
Every DOI below has been checked. Where I could not confirm a figure from the paper itself, I have left the figure out rather than guess.
📃 Erford and colleagues (2025) 193 studies. Internal consistency .85 to .92; no structural validity or test-retest studies in 25 years. https://doi.org/10.1002/jcad.70006
📃 Momm and colleagues (2015) 142 employee, peer and supervisor triads. Emotion recognition linked to income via political skill. https://doi.org/10.1002/job.1975
📃 Harrison and colleagues (2019) A language-based Big Five measure applied to more than 3,000 chief executives; traits shaped strategic change. https://doi.org/10.1002/smj.3023
📃 Savitsky and colleagues (2011) The closeness-communication bias: more egocentrism in communication between friends than strangers. https://doi.org/10.1016/j.jesp.2010.09.005
📃 Ma-Kellams and Blascovich (2013) An ironic effect of financial incentive on empathic accuracy. https://doi.org/10.1016/j.jesp.2012.08.014
📃 Van Der Wege and colleagues (2021) Familiarity, group membership and shared experience fed overconfidence in being understood. https://doi.org/10.1016/j.jesp.2020.104097
📃 Letzring (2008) Good judges were socially skilled, agreeable and well adjusted, and drew out revealing behaviour. https://doi.org/10.1016/j.jrp.2007.12.003
📃 Haslam, Holland and Kuppens (2012) 177 articles. Apparent categories fell from 38.9% to about 14% once method was accounted for. https://doi.org/10.1017/s0033291711001966
📃 Haslam and colleagues (2020) 317 findings. Dimensional results outnumbered categorical ones five to one. https://doi.org/10.1017/s003329172000183x
📃 Ames and Kammrath (2004) Narcissism, not actual competence, predicted how good people thought they were at reading others. https://doi.org/10.1023/b:jonb.0000039649.20015.0e
📃 Fiske (1993) Power and stereotyping reinforce each other through where attention goes. https://doi.org/10.1037/0003-066x.48.6.621
📃 Judge and colleagues (2002) 222 correlations, 73 samples. Extraversion .31, conscientiousness .28, openness .24, emotional stability .24, agreeableness .08 with leadership. https://doi.org/10.1037/0021-9010.87.4.765
📃 Tett and Burnett (2003) Trait activation: situations at work call personality out, or hold it back. https://doi.org/10.1037/0021-9010.88.3.500
📃 Peterson and colleagues (2003) 17 chief executives. Their personality shaped top team dynamics, which shaped performance. https://doi.org/10.1037/0021-9010.88.5.795
📃 Nisbett and Wilson (1977) The halo effect: a warm or cold manner changed ratings of accent and appearance, unnoticed. https://doi.org/10.1037/0022-3514.35.4.250
📃 Snyder and Swann (1978) People tested hypotheses about others by seeking confirmation, and drew confirming behaviour out. https://doi.org/10.1037/0022-3514.36.11.1202
📃 Neuberg and Fiske (1987) Depending on someone led people to attend to that person's individual attributes, not a category. https://doi.org/10.1037/0022-3514.53.3.431
📃 Ickes and colleagues (1990) 38 pairs of strangers. Empathic accuracy could be measured reliably, and self-rated empathic skill did not predict it. https://doi.org/10.1037/0022-3514.59.4.730
📃 Stinson and Ickes (1992) Friends read each other better than strangers did, because they knew each other's lives. https://doi.org/10.1037/0022-3514.62.5.787
📃 Shoda, Mischel and Wright (1994) Children's behaviour across situations formed stable, distinctive if-then profiles. https://doi.org/10.1037/0022-3514.67.4.674
📃 Webster and Kruglanski (1994) The need for cognitive closure: how much someone wants a firm answer quickly. https://doi.org/10.1037/0022-3514.67.6.1049
📃 Swann and Gill (1997) Longer, closer relationships raised confidence in impressions of a partner without reliably raising accuracy. https://doi.org/10.1037/0022-3514.73.4.747
📃 Gilovich, Savitsky and Medvec (1998) The illusion of transparency: people overestimate how visible their inner states are. https://doi.org/10.1037/0022-3514.75.2.332
📃 Gesn and Ickes (1999) Accuracy depended mainly on what was said; context sometimes helped and sometimes misled. https://doi.org/10.1037/0022-3514.77.4.746
📃 Fleeson (2001) Traits as spreads of states; behaviour varied widely around each person's average. https://doi.org/10.1037/0022-3514.80.6.1011
📃 Simpson, Oriña and Ickes (2003) Reading a partner accurately when their thoughts were threatening lowered the reader's closeness. https://doi.org/10.1037/0022-3514.85.5.881
📃 Thomas and Fletcher (2003) Mind-reading accuracy rose with acquaintance and with prior disclosure about the problem at hand. https://doi.org/10.1037/0022-3514.85.6.1079
📃 Gill and Swann (2004) Close partners were most accurate about what mattered to the relationship, and that accuracy went with harmony. https://doi.org/10.1037/0022-3514.86.3.405
📃 Borkenau and colleagues (2004) 600 people, 15 short episodes. Thin slices tracked well-informed views; gains flattened after six. https://doi.org/10.1037/0022-3514.86.4.599
📃 Kammrath, Mendoza-Denton and Mischel (2005) People link traits to if-then patterns and use them to infer motives. https://doi.org/10.1037/0022-3514.88.4.605
📃 Letzring, Wells and Funder (2006) More information, and better information, both raised the accuracy of personality judgements. https://doi.org/10.1037/0022-3514.91.1.111
📃 Biesanz, West and Millevoi (2007) With acquaintance, the distinctive picture of a person improved and the generic one faded. https://doi.org/10.1037/0022-3514.92.1.119
📃 Kenny and Albright (1987) Accuracy about a person in general is different from accuracy about how they behave with you. https://doi.org/10.1037/0033-2909.102.3.390
📃 Kenny and DePaulo (1993) 8 studies, 569 people. People infer how others see them mostly from how they see themselves. https://doi.org/10.1037/0033-2909.114.1.145
📃 Kenny and colleagues (1994) 32 studies. Agreement between judges ranged from zero to about .3 and did not rise with acquaintance. https://doi.org/10.1037/0033-2909.116.2.245
📃 Gilbert and Malone (1995) The correspondence bias: putting behaviour down to character when the situation explains it. https://doi.org/10.1037/0033-2909.117.1.21
📃 Roberts and DelVecchio (2000) 152 longitudinal studies. Trait consistency rose from .31 in childhood to a plateau around .74. https://doi.org/10.1037/0033-2909.126.1.3
📃 Mischel and Shoda (1995) Personality as stable if-then patterns across situations, not a fixed average. https://doi.org/10.1037/0033-295x.102.2.246
📃 Funder (1995) The realistic accuracy model: a read needs relevant behaviour, made available, detected and used well. https://doi.org/10.1037/0033-295x.102.4.652
📃 Pittenger (2005) Caution warranted when making inferences from the four-letter type formula. https://doi.org/10.1037/1065-9293.57.3.210
📃 Hogan and Kaiser (2005) Review. Personality predicts leadership, and leadership is about team performance. https://doi.org/10.1037/1089-2680.9.2.169
📃 Hall and Schmid Mast (2007) Words carried the most accuracy about thoughts and feelings, then the voice, then the face. https://doi.org/10.1037/1528-3542.7.2.438
📃 Schmid Mast, Jonas and Hall (2009) Four studies. Feeling powerful raised interpersonal sensitivity, especially with an empathic style. https://doi.org/10.1037/a0016234
📃 Vazire (2010) 165 people. The self knew neuroticism best, friends knew intellect best, everyone saw extraversion. https://doi.org/10.1037/a0017908
📃 Connelly and Ones (2010) 44,178 people rated by others. Closeness, more than contact, improved accuracy for hidden traits. https://doi.org/10.1037/a0021212
📃 Oh, Wang and Mount (2011) Observer ratings of personality predicted job performance better than self-reports. https://doi.org/10.1037/a0021832
📃 Carlson, Vazire and Furr (2011) People do have some real insight into how others see them. https://doi.org/10.1037/a0024297
📃 Goldstein, Vezich and Shapiro (2014) Six experiments. Feeling that someone had taken your perspective raised liking and helping. https://doi.org/10.1037/a0036395
📃 Cronbach (1955) How scores on understanding others are shaped by assumed similarity. https://doi.org/10.1037/h0044919
📃 Forer (1949) The classroom demonstration: everyone accepted the same sketch as a description of themselves. https://doi.org/10.1037/h0059240
📃 Murphy and Lilienfeld (2019) 85 studies, 14,327 people. Self-reported ability to read others explained about 1% of measured ability. https://doi.org/10.1037/pas0000732
📃 Wendt and colleagues (2024) 700 people, preregistered. Self-rated mind-reading correlated .05 with measured mind-reading. https://doi.org/10.1037/pas0001310
📃 Ma-Kellams and Lerner (2016) Executive-level professionals. Intuitive thinkers read emotions less accurately; inducing systematic thought helped. https://doi.org/10.1037/pspi0000063
📃 Ginosar Yaari and colleagues (2025) 2,225 people. We see our own values in the people close to us. https://doi.org/10.1037/pspi0000482
📃 Rogers and Biesanz (2019) Four large samples. Good judges exist, and show most clearly when the person being read is easy to read. https://doi.org/10.1037/pspp0000197
📃 Todd and Tamir (2024) Review of egocentric mentalizing: our own mind is the starting point, and we adjust too little. https://doi.org/10.1038/s44159-024-00277-1
📃 Blanch-Hartigan, Andrzejewski and Hill (2012) Meta-analysis. Training improved person perception; practice and feedback beat instruction alone. https://doi.org/10.1080/01973533.2012.728122
📃 McCrae and Costa (1989) 468 adults. The indicator measures four dimensions, not distinct types, overlapping four of the Big Five. https://doi.org/10.1111/j.1467-6494.1989.tb00759.x
📃 Howland and Rafaeli (2010) Three weeks of diaries. Partners read negative moods more accurately than positive ones. https://doi.org/10.1111/j.1467-6494.2010.00657.x
📃 Galinsky and colleagues (2006) Feeling powerful reduced perspective taking and accuracy in reading emotional expressions. https://doi.org/10.1111/j.1467-9280.2006.01824.x
📃 Zaki, Bolger and Ochsner (2008) Perceivers' empathy predicted accuracy only when the person they were reading was expressive. https://doi.org/10.1111/j.1467-9280.2008.02099.x
📃 Ickes, Gesn and Graham (2000) 15 studies. Women outperformed men only when empathy was made salient: motivation, not ability. https://doi.org/10.1111/j.1475-6811.2000.tb00006.x
📃 Kaplan, Klebanov and Sorensen (2012) Candidates for chief executive varied on general ability and on interpersonal versus execution skills. https://doi.org/10.1111/j.1540-6261.2012.01739.x
📃 Byron, Terranova and Nowicki (2007) Salespeople who read nonverbal emotion better earned more and sold more, without rating themselves higher. https://doi.org/10.1111/j.1559-1816.2007.00272.x
📃 Boyle (1995) Psychometric limitations of the type indicator; routine use not recommended. https://doi.org/10.1111/j.1742-9544.1995.tb01750.x
📃 Barrick and Mount (1991) Conscientiousness predicted performance across jobs; extraversion for managers and sales. https://doi.org/10.1111/j.1744-6570.1991.tb00688.x
📃 Cameron and Vorauer (2008) Review. Closeness and self-awareness make people overestimate their transparency, with real costs. https://doi.org/10.1111/j.1751-9004.2008.00096.x
📃 Bleidorn, Hopwood and Lucas (2018) Review. Some, still preliminary, evidence that major life events can change traits. https://doi.org/10.1111/jopy.12286
📃 Todorov and colleagues (2015) Review. Judgements from faces are fast and consequential; their accuracy has been greatly overstated. https://doi.org/10.1146/annurev-psych-113011-143831
📃 Capraro and Capraro (2002) Reliability generalisation. The indicator's scores were generally reliable, with variation. https://doi.org/10.1177/0013164402062004004
📃 Scott and Bruce (1995) Five decision-making styles: rational, intuitive, dependent, avoidant and spontaneous. https://doi.org/10.1177/0013164495055005017
📃 Klein and Hodges (2001) Paying for accuracy improved men's and women's empathic accuracy and removed the gap. https://doi.org/10.1177/0146167201276007
📃 Meyer, Dalal and Hermida (2010) Strong situations restrict how much personality shows. https://doi.org/10.1177/0149206309349309
📃 Soto (2019) 78 trait-outcome links replicated. 87% held, at about 77% of their original strength. https://doi.org/10.1177/0956797619831612
📃 He and Côté (2023) 1,126 students and 2,205 informants. Real empathic accuracy went with better relationships; believing you had it did not. https://doi.org/10.1177/09567976231185127
📃 Jussim, Crawford and Rubinstein (2015) Review arguing stereotypes are often more accurate than assumed. https://doi.org/10.1177/0963721415605257
📃 Schmid Mast and Hall (2018) Interpersonal accuracy works through adaptability rather than any one behaviour. https://doi.org/10.1177/0963721418758437
📃 Kenny and West (2010) Meta-analysis of 24 studies of agreement and similarity in self and other perception. https://doi.org/10.1177/1088868309353414
📃 Human and Biesanz (2013) Review of what makes a person easy to read: adjustment, status and socialisation. https://doi.org/10.1177/1088868313495593
📃 Human and colleagues (2013) Accurate first impressions of new classmates predicted more interaction and liking later. https://doi.org/10.1177/1948550612463735
📃 Gunaydin, Selcuk and Zayas (2017) Liking from a photograph predicted liking a month later after meeting, partly through warmer behaviour. https://doi.org/10.1177/1948550616662123
📃 Dickson and Kelly (1985) Review of the Barnum effect: vague, favourable descriptions feel personal. https://doi.org/10.2466/pr0.1985.57.2.367
📃 Pittenger (1993) Insufficient evidence to support the claims made for the type indicator's usefulness. https://doi.org/10.3102/00346543063004467
📃 Rubin, Munz and Bommer (2005) 145 managers. Emotion recognition predicted transformational leadership behaviour. https://doi.org/10.5465/amj.2005.18803926
📃 Durán and Fernández-Dols (2021) Meta-analysis. The classic emotions co-occurred with their predicted faces at only .13 to .30. https://doi.org/10.1037/emo0001015
📃 Elfenbein and Ambady (2002) Emotion recognition was above chance everywhere, and best within a shared culture. https://doi.org/10.1037/0033-2909.128.2.203
📃 Nadkarni and Herrmann (2010) 195 firms. Chief executive personality drove how quickly the firm adapted. https://doi.org/10.5465/amj.2010.54533196
📃 Hambrick and Mason (1984) The upper echelons view: organisations partly reflect their top managers. https://doi.org/10.5465/amr.1984.4277628
📃 Eyal and Epley (2010) Matching the level of detail at which the other person thinks made mind reading more accurate. https://doi.org/10.1177/0956797610367754
📃 Kruger and colleagues (2005) Five experiments. People overestimated how well their tone came across in email. https://doi.org/10.1037/0022-3514.89.6.925
P.P.S. Why Do I Even Have the Nerve to Write This?
Fair question.
I am someone who has spent years working alongside remarkable executives, navigating chaos, translating vision into structure and figuring things out, sometimes beautifully and sometimes the very hard way. I have had the privilege of being the steady right hand to leaders who move fast, take risks and expect a lot. That environment has shaped me more than any classroom ever could.
I also happen to have an MBA, and I am studying psychology because people fascinate me. How we work. Why we disconnect. What actually holds us together when the pressure rises.
But really, none of that is the point.
I am here because I have lived the erosion and rebuilt from it. I have seen the cost of carrying too much and the relief that arrives when you finally stop disappearing inside your own competence.
So I share what I have learned in case it helps someone else. Take what is useful. Leave the rest. And remember this one truth that professionals like you often forget:
You are probably doing far better than you give yourself credit for.
Comments