Psychologist logo
Personality and self

The paradox of knowing

Why do we have greater insight into others than ourselves? David Dunning outlines some intriguing research.

26 June 2013

To know others is wisdom, to know one’s self is enlightenment.
Chinese philosopher Lao Tzu

For the past twenty-odd years, the main discovery in my lab has been finding out just how unenlightened people are, at least in the terms that Lao Tzu put it. People appear to harbour many and frequent false beliefs about their own competence, character, place in the social world, and future (Dunning, 2005; Dunning et al., 2004). If ‘knowing yourself’ is a task that many philosophers and social commentators – from both Western and Eastern traditions – have exhorted people to accomplish, it appears that very few are taking the advice seriously enough to succeed.

But here is the rub. Although people may not possess much enlightenment, according to Lao Tzu’s criteria, they do instead seem to display a lot of wisdom. At least when it comes to making predictions about the future, people achieve more accuracy forecasting what their peers will do than what they themselves will do. Through their predictions, they seem to possess a rough but valid wisdom about the general dynamics of human nature and how it is reflected in people’s actions. They just fail to display the same sagacity when it comes to understanding their own personal dynamics. As psychologists, they appear to be much better social psychologists than self-psychologists.

The ‘holier-than-thou’ phenomenon
in behavioural prediction perhaps best illustrates this paradox of greater insight into other people than the self. The phenomenon is defined as people predicting they are far more likely to engage in socially desirable acts than their peers. Across several studies, we have asked people to forecast how they will behave in situations that have an ethical, civic or altruistic tone. For example, we ask whether they will donate to charity, or cooperate with another person in an experiment, or vote in an upcoming election. We also ask them the likelihood that their peers will do the same. Consistently, we find that respondents claim that they are much more likely to act in a socially desirable way than their peers are (Balcetis & Dunning, 2008, 2013; Epley & Dunning, 2000, 2006).

But here is the key twist: We then expose an equivalent set of respondents to the actual situation, to see which prediction – self or peer – better anticipates the true rate at which people ‘do the right thing’. Do self-predictions better anticipate the rate that people act in desirable ways, with people, thus, showing undue cynicism about the character of their peers? Or do peer predictions prove more accurate, demonstrating that people believe too much in their better selves?In our studies we find that people’s peer predictions are the more accurate ones. Self-predictions, in contrast, are wildly optimistic. For example, in one study, a full 90 per cent of students in a large-lecture psychology class eligible to vote in an upcoming US presidential election said that they would. They then provided another student with some relevant information about themselves, such as how interested they were in the election and how pleased would they be if their favoured candidate won. Peers given such information predicted that only 67 per cent of respondents would vote. Actual voting rate among those respondents when the election arrived: 61 per cent (Epley & Dunning, 2000, Study 2).

Time and again we have seen such a pattern. For example, 83 per cent of students forecast that they would buy a daffodil for charity in an upcoming drive for the American Cancer Society, but that only 56 per cent of their peers would. When we check back, we found that only 43 per cent had done so (Epley & Dunning, 2000, Study 1). In a Prisoner’s Dilemma game played in the lab, 84 per cent of participants said they would cooperate rather than betray their partner, but that only 64 per cent would do likewise. The actual cooperation rate was 61 per cent (Epley & Dunning, Study 2).

Accuracy as correlation
But wait, a careful reader might say. People might prove overconfident about their own behaviour, but surely they know more about themselves than other people do. This accuracy just reveals itself in a different way. Namely, if we look instead at the correlation between people’s predictions and their actions, we might find a stronger relationship for self-predictions than for peers. More specifically, people may overpredict the chance that they will vote. But those who say they will vote will still be much more likely to vote than those who say they will not. Forecasts from peers will fail to separate voters from nonvoters so successfully.

This assertion is plausible, but it surprisingly fails empirical test. When we look at accuracy from a correlational perspective, we find that peers at least equal overall the accuracy rates of those making self-predictions (see also Spain et al., 2000; Vazire & Mehl, 2008). In one of our voting studies, peers who received just five scant pieces of information about another person’s view of an upcoming election predicted that person just as well (r = .48) as did people predicting their own actions (r = .51) in correlational terms. Other researchers report similar findings: All it takes is a few pieces of information for
a peer to achieve accuracy rates that equal the self. The behaviour can be a performance in an upcoming exam (Helzer & Dunning, 2012) or performance on IQ tests (Borkenau & Liebler, 1993).

And, if the action is one that people find significant, and if peers are familiar with the person in question, then peer prediction begins to outdo self-prediction. Roommates and parents, for example, outpredict how long a person’s college romance will last, relative to self-prediction (MacDonald & Ross, 1999). Ratings of supervisors and peers outclass self-ratings in predicting how well surgical residents will do on their final surgical exams (Riscucci et al., 1989). Ratings of peers do better at predicting who will receive a promotion in the Navy early relative to self-impressions (Bass & Yammarino, 1991).

Misguided exceptionalism

Taken together, all this research suggests that people tend to possess useful insight when it comes to understanding human nature. But this research also suggests that people fail to apply this wisdom to the self. In a sense, people exempt themselves from whatever valid psychological understanding they have about their friends and contemporaries. Instead, they tend to think of themselves as special, as responding to a different psychological dynamic. The rules that govern other people’s psychology fail to apply to them. We have come to call this tendency misguided exceptionalism.

What is it about their understanding of other people that respondents exempt themselves from? We contend, with data, that people recognise that others tend to be constrained in what they do. There are forces, both internal and external to the individual, which are out of their control but that influence how they behave. The smell of freshly-baked chocolate chip cookies does break people’s willpower.

The opinions of the crowd place pressures on other people to conform.
But these constraints are for other people. When it comes to our own behaviour, we tend to emphasise instead our own agency, the force of our own character, and what we aspire, intend or plan to do. Relative to others, we believe that our actions are largely a product of our own intentions, aspirations and free will (Buehler et al., 1994; Critcher & Dunning, 2013; Koehler & Poon, 2006; Kruger & Gilovich, 2004; Peetz & Buehler, 2009). We consider ourselves free agents generally immune to the constraints that dictate other people’s actions.

Much recent empirical work reveals this differential emphasis for the self. People think their futures are more wide-open and unpredictable, and that their intentions and desires will be more important authors of their futures than similar intentions and desires will be for other people (Pronin & Kugler, 2010). When predicting their own exam performance, people emphasise (actually, too much, it turns out) their aspiration level, that is, the score they are working to achieve (Helzer & Dunning, 2012), but they emphasise instead a person’s past achievement (appropriately, it turns out) in predictions of others. College students consider their future potential – or, rather, the person they are aiming to be – to be a bigger part of themselves than it is in other people (Williams & Gilovich, 2008; Williams et al., 2012). People predicting who will give to charity consider the prediction to be one about a person’s character and attitudes – that is, until they confront a chance to give themselves, in which case they switch to emphasising situational factors in their accounts of giving (Balcetis & Dunning, 2008).

Misunderstanding situations
Ultimately, this misguided exceptionalism and overemphasis on individual agency means that people fail to apply an accurate understanding of human nature to themselves, one that would make their predictions more accurate. People, for example, are surprisingly good at understanding how situational circumstances influence people’s behaviour. In one study, we described a ‘bystander apathy’ study to students. Students were shown an experiment in which a research assistant accidentally spilled a box of jigsaw puzzle pieces. These students were then asked the likelihood that they would help pick the pieces up relative to the percentage of other students who would help. Of key importance, participants were shown two variations of this basic situation – one in which they were alone versus one in which they were sitting in a group of three people.

Those familiar with social psychology will recognise that people are more likely to help when they are alone rather than in a group (Latané & Darley, 1970). In the group, people are seized by the inertia of not knowing immediately whether to help, and thus taking their cue to do nothing based on the fact that everyone else, lost in the same indecision, ends up doing nothing, too. But would our participants show insight into this principle? Not according to their self-predictions. Participants stated that they would be roughly 90 per cent likely to help either alone or in the group. They did, though, concede that other people would be influenced, and that the rate of helping would go down 22 per cent (from 72 per cent to 50 per cent) among other people by introducing the group. Of key import, when we ran the study for real, we found that placing people in a group had a 27 per cent impact (from 50 per cent down to 23 per cent) on actual behaviour. Again, peer predictions largely anticipated this impact. Self-predictions did not (Balcetis & Dunning, 2013).
This belief that self-behaviour ‘floats’ above the impact of situational circumstances and constraints can lead people to forgo decisions that would actually help them. Consider the task of staying within a monthly budget. In one study, participants were offered a service that would provide them with savings tips plus a constant monitoring of their finance. For themselves, participants felt the service would be superfluous. It would have almost zero impact on their ability to achieve their budget goals. What mattered for them instead was the strength of their intentions to save money (Koehler et al., 2011).

But, in reality, a random sample of participants assigned to the service was roughly 11 per cent more likely to reach their budget goals. And, a group of participants asked to judge the impact of the service on other people estimated that the service would matter; that others would be 17 per cent more likely to reach their goals. Again, predictions about others better reflected reality than predictions about the self, in that people could recognise the impact of an important situational aid on others, but felt they themselves were immune to those influences (Koehler et al., 2011).

Cultural influences
This overemphasis on the self’s agency suggests possible cultural differences in the holier-than-thou effect. And, indeed, such cultural differences arise. It is the individualist cultures of Western Europe and North America that emphasise autonomy, agency and the imposition of will onto the environment (Fiske et al., 1998; Markus & Kitayama, 1991). Far Eastern cultures, such as Japan, emphasise instead interdependence, social roles and group harmony – that is, social constraints on the self. Might those cultures, thus, be relatively immune to the ‘holier’ phenomenon?

Across several studies, we have found that people from collectivist cultures display much less self-error than did those from individualist ones. For example, young children attending a summer school on Mallorca were asked how many candies they would donate to other children if they were asked, as well as how many candies other children on average would donate.

A week later, the children were actually asked to donate. Children from more individualist countries (e.g. Britain) donated many fewer candies than they had predicted, but those from more collectivist countries (e.g. Spain) donated on average just as many as they had predicted. Both groups were accurate in their predictions about their peers (Balcetis et al., 2008).

Does the self have any advantage?
Extant psychological research, however, does suggest one area where this general story about self- and social insight will reverse. People may be wiser when it comes to predicting the public and observable actions of others rather than self, but they do appear to have privileged insight into aspects of the self that are not available for other people to view. People know that below the surface of their public appearance is a private individual who feels doubt, anxiety, inhibition and ambivalence that he or she may not let wholly come to the surface (Spain et al., 2000; Vazire, 2010; Vazire & Carlson, 2010, 2011). Of course, this individual does not see this roiling interior life in others.

As a consequence, people may lack awareness that the what’s inside themselves is similarly churning and stirring within others. Thus, for example, people often consider themselves more shy, self-critical, and indecisive than other people (Miller & McFarland, 1987). College students harbour reservations about excessive drinking, but not recognising that others also feel this same reluctance, they go along with the crowd to excess on a Saturday night (Prentice & Miller, 1993). In a similar vein, college students harbour much more discomfort about casual sex than they believe their peers do, with each sex overestimating the comfort level of the other sex when it comes to ‘hooking up’ (Lambert et al., 2003).

Concluding remarks
Thus, current psychological research suggests that people may be wise, at least when it comes to understanding and anticipating other people, but they stand in the way of letting this wisdom lead to their own enlightenment. However, if research reveals this problem, it also suggests a potential solution to it. What we presume about other people’s behaviour and futures is likely a valuable indicator of what awaits us in the same situation – and may be much better indicator of our future than any scenario we are spinning directly about ourselves. When predictions matter, we should not spend a great deal of time predicting what we think we will do. Instead, we should ask what other people are likely to do. Or, we should hand the prediction of our own future over to another person who knows a little about us.

Whatever we do, we should note that perhaps we are, indeed, uniquely special individuals, but that it is too easy to overemphasise that fact. In anticipating the future, we should be mindful of the continuity that lies between our self-nature and the nature of others. It is in recognising this continuity that we realise the path that leads to our wisdom may be
a pretty good path to our enlightenment, too. At the very least, that thought does remind one of another Chinese proverb that has survived the centuries, perhaps best indicating its worth – that to know what lies for us along the road ahead, we should be sure to ask those coming back.

David Dunning
is at the Department of Psychology, Uris Hall, Cornell University, Ithaca, New York
[email protected]


Balcetis, E. & Dunning, D. (2008). A mile in moccasins: How situational experience reduces dispositionism in social judgment. Personality and Social Psychology Bulletin, 34, 102–114.
Balcetis, E. & Dunning, D. (2013). Considering the situation: Why people are better social psychologists than self-psychologists. Self and Identity, 12, 1–15.
Balcetis, E., Dunning, D. & Miller, R.L. (2008). Do collectivists ‘know themselves’ better than individualists? Journal of Personality and Social Psychology, 95, 1252–1267.
Bass, B.M. & Yammarino, F.J. (1991). Congruence of self and others’ leadership ratings of Naval officers for understanding successful performance. Applied Psychology, 40, 437–454.
Buehler, R., Griffin, D. & Ross, M. (1994). Exploring the ‘planning fallacy’. Journal of Personality and Social Psychology, 67, 366–381.
Borkenau, P. & Liebler, A. (1993). Convergence of stranger ratings of personality and intelligence with self-ratings, partner ratings, and measured intelligence. Journal of Personality and Social Psychology, 65, 546–553.
Critcher, C.R. & Dunning, D. (2013). Predicting persons’ goodness versus a person’s goodness: Forecasts diverge for populations versus individuals. Journal of Personality and Social Psychology, 104, 28–44.
Dunning, D. (2005). Self-insight: Roadblocks and detours on the path to knowing thyself. New York: Psychology Press.
Dunning, D., Heath, C. & Suls, J. (2004). Flawed self-assessment: Implications for health, education, and the workplace. Psychological Science in the Public Interest, 5, 71–106.
Epley, N. & Dunning, D. (2000). Feeling ‘holier than thou’: Are self-serving assessments produced by errors in self or social prediction? Journal of Personality and Social Psychology, 79, 861–875.
Epley, N. & Dunning, D. (2006). The mixed blessings of self-knowledge in behavioral prediction. Personality and Social Psychology Bulletin, 32, 641–655.
Fiske, A., Kitayama, S., Markus, H.R. & Nisbett, R.E. (1998). The cultural matrix of social psychology. In D. Gilbert, S. Fiske & G. Lindzey (Eds.) The handbook of social psychology (4th edn, pp.915–981). San Francisco: McGraw-Hill.
Helzer, E.G. & Dunning, D. (2012). Why and when peer prediction is superior to self-prediction. Journal of Personality and Social Psychology, 103, 38–53.
Koehler, D.J. & Poon, C.S.K. (2006). Self-predictions overweight the strength of current intentions. Journal of Experimental Social Psychology, 42, 517–524.
Koehler, D.J., White, R.J. & John, L.K. (2011). Good intentions, optimistic self-predictions, and missed opportunities. Social Psychological and Personality Science, 2, 90–96.
Kruger, J. & Gilovich, T. (2004). Actions and intentions in self-assessments: The road to self-enhancement is paved with good intentions. Personality and Social Psychology Bulletin, 30, 328–339.
Lambert, T.A., Kahn, A.S. & Apple, K.J. (2003). Pluralistic ignorance and hooking up. Journal of Sex Research, 40, 129–133.
Latané, B. & Darley, J. (1970). The unresponsive bystander: Why doesn't he help? New York: Appleton-Century-Crofts.
MacDonald, T.K. & Ross, M. (1999). Assessing the accuracy of predictions about dating relationships. Personality and Social Psychology Bulletin, 25, 1417–1429.
Markus, H.R. & Kitayama, S. (1991). Culture and the self. Psychological Review, 98, 224–253.
Miller, D.T. & McFarland, C. (1987). Pluralistic ignorance: When similarity is interpreted as dissimilarity. Journal of Personality and Social Psychology, 53, 298–305.
Peetz, J. & Buehler, R. (2009). Is there a budget fallacy? Personality and Social Psychology Bulletin, 35, 1579–1591.
Prentice, D.A. & Miller, D.T. (1993). Pluralistic ignorance and alcohol use on campus: Some consequences of misperceiving the social norm. Journal of Personality and Social Psychology, 64, 243–356.
Pronin, E. & Kugler, M.B. (2010). People believe they have more free will than others. Proceedings in the National Academy of Sciences, 107, 22469–22474.
Risucci, D.A., Tortolano, A.J. & Ward, R.J. (1989). Ratings of surgical residents by self, supervisors and peers. Surgical Gynecology and Obstetrics, 169, 519–526.
Spain, J.S., Eaton, L.G. & Funder, D.C. (2000). Perspectives on personality. Journal of Personality, 68, 837–867.
Vazire, S. (2010). Who knows what about a person? Journal of Personality and Social Psychology, 98, 281–300.
Vazire, S. & Carlson, E.N. (2010). Self-knowledge of personality. Social and Personality Psychology Compass, 4, 605–620.
Vazire, S. & Carlson, E.N. (2011). Others sometimes know us better than we know ourselves. Current Directions in Psychological Science, 20, 104–108.
Vazire, S., & Mehl, M.R. (2008). Knowing me, knowing you. Journal of Personality and Social Psychology, 95, 1202–1216.
Williams, E.F. & Gilovich, T. (2008). Conceptions of the self and others across time. Personality and Social Psychology Bulletin, 34, 1037–1046.
Williams, E., Gilovich, T. & Dunning, D. (2012). Being all that you can be. Personality and Social Psychology Bulletin, 38, 143–154.