Strangers to ourselves?
David Dunning with some fascinating studies into how and why we are deluded about the self.
18 October 2006
Ninety per cent of the world’s woe comes from people not knowing themselves, their abilities, their frailties, and even their real virtues. Most of us go almost all the way through life as complete strangers to ourselves.
Sidney J. Harris, American Journalist, 1917–1986
For centuries, philosophers, theologians, psychologists, and social commentators have stressed the critical importance of accurate self-knowledge for attaining success and fulfillment. Indeed, how can ‘knowing thyself’ be anything less than invaluable?
A barrister, to ensure winning their case, needs to know where their knowledge of relevant law ends and the need to dig into more research begins. A teenage driver, before grabbing the keys to their parent’s car, better have a good sense of their ability to handle bad road conditions before driving out into a wintry storm.
Given this, it is perhaps disconcerting to note that Sydney Harris’ dour assessment of self-insight, although certainly extreme, lies closer to the portrait painted by psychological research than one might initially believe. Self-perceptions of competence and character are frequently filled with heavy doses of bias, misconception, and illusion. Such perceptions carry often only a modest relationship to reality, if any relationship emerges at all.
A tenuous relation
Many studies reveal the frequently tenuous association between the perception and reality of self. Sure, self-perceptions carry some value, in that they tend to be positively related to performance. But meta-analyses suggest that this relationship is not very strong, on average producing a correlation around .29 (Mabe & West, 1982; see also Dunning, 2005; Dunning et al., 2004, for a review).
In addition, in some areas in which one would like the correlation to be larger, it simply isn’t. Consider professionals responsible for public health and safety. Obstetrics/gynaecology residents just about to finish their schooling accurately predict their final grades for verbal and written skills to a modest degree, but not their grades for general knowledge, technical expertise, or clinical problem-solving skills (Weiss et al., 2005). Family practitioners rating their knowledge about thyroid disorders fail to show any insight into their actual level of knowledge (Tracey et al., 1997). Medical residents’ impressions of their communication skills show little relationship with impressions held by patients or supervisors, although the impressions of patients and supervisors correlate rather highly (Millis et al., 2002). And in an anxious age in which discussions of terrorism dominate public life, one would like public health workers to be knowledgeable about how to respond to an emergency. However, the correlation between perceived expertise about local response plans and actual knowledge hovers around .34 (Kerby et al., 2005).
In fact, at times, other people assess our competence and prospects better than we do ourselves. Roommate ratings better predict which college romances survive than do self-impressions of the romance (MacDonald & Ross, 1999). Peer ratings amongst junior doctors strongly predict who will do well on a surgical skills exam; self-ratings do not (Risucci et al., 1989).
In predictions of who will receive early promotion among Naval officers, peer ratings of leadership ability prove to be a more accurate indicator than self-ratings (Bass & Yammarino, 1991). And after watching a mere 90-second videotape of a target person reading a weather report, people provide ratings of intelligence that predict that target’s scores on intellectual tests just as well as the target’s self-ratings (Borkenau & Liebler, 1993).
Overconfidence in self
Self-assessments are not just inaccurate, they tend to be flawed in a particular direction. Franklin Roosevelt once quipped to an acquaintance that ‘the ablest man I ever met is the one you think you are.’ It appears that the rest of us have similarly immodest and unrealistic views of ourselves.
This overconfidence in self shows up in a myriad of ways. People, for example, overpredict the likelihood that they will perform generous, kind, and ethical acts. People overestimate the odds that they will buy a flower for charity, vote for President, maintain a successful romantic relationship, volunteer for an unpleasant lab experiment so that a 10-year-old girl will not have to, and cooperate with another person when money is at stake (Epley & Dunning, 2000, 2006). Curiously, people mispredict themselves even though they are roughly accurate in predicting the rate at which their peers perform these same actions.
People also overestimate how quickly they can get tasks done, a phenomenon known as the planning fallacy. College students completing a senior-year thesis in one study tend to finish their thesis three weeks later than their ‘most realistic’ estimate, and one week later than their ‘worst case scenario’. People filing income tax returns commonly mail in their forms a week later than they thought (Buehler et al., 2002). Indeed, at Microsoft, the exhausting scramble that software developers put themselves through to meet their overly optimistic deadlines has proven to be a central source of stress and burnout (Cusumano & Selby, 1995).
But perhaps the most pervasive demonstration of this overconfidence comes from what people think of their competence and character relative to other people. On average, people think of themselves as anything but average. The average person claims to be more disciplined, idealistic, socially skilled, a better driver, good at leadership, and healthier than the average person. Mathematically, this cannot be right – the average person cannot be above average. But many surveys (with some intriguing exceptions) have shown that people believe they excel among their contemporaries. Ironically, people also claim to be better than most other people at producing unbiased and realistic self-estimates (for reviews, see Dunning, 2005; Dunning et al., 2004).
Good or bad?
I think two central points should be made about the flawed nature on people’s self-assessments. First, one could wonder whether the tendency to overrate one’s
self is costly or beneficial. After all, in contemporary Western culture, it is perfectly acceptable to be a confident and optimistic person. Perhaps overdoing it is not so bad.
Indeed, work by Shelley Taylor and others has shown some contexts where optimism may be advantageous (e.g. Taylor & Brown, 1988). When people face some of the most difficult challenges of human experience – such as suffering from cancer – it may be helpful to overestimate one’s self, even to an unrealistic degree, although this thesis still produces contentious debate in the psychological literature.
However, for the rest of us, there are contexts where self-overestimation can be quite costly. Consider the business world. Companies trying to acquire other firms typically offer to pay 41 per cent more for the target firm’s stock than its current price. Essentially, this is a statement that they believe they can run the target firm much more profitably than its current management. However, on the days these offers are announced, the combined stock price of the two companies usually falls – suggesting the outside world, voting with money, does not endorse the acquiring firm’s rosy opinion of its management prowess (Malmendier & Tate, 2003).
Consider, in addition, patients being treated for high blood pressure. Medical experts agree that there are no overt symptoms that reliably alert people to episodes of high blood pressure. But one study showed that 92 per cent of patients claimed that they could tell when their blood pressure was up (with over 60 per cent confidentially asking the interviewer not to reveal this opinion to their doctor). These misplaced beliefs about self-diagnoses matter. Patients tended to take their prescribed medication only when they thought it treated the ‘symptoms’ they were paying attention to, instead of taking their medication as their doctor prescribed (Meyer et al., 1985).
Vaulted self-views also underlie risk-taking behaviors that have an impact on health. Teenage girls who rate their knowledge of birth control methods highly, independent of actual knowledge, are more likely to become pregnant than those who rate their knowledge more negatively (Jaccard et al., 2005). Of elderly drivers referred for a driving evaluation, those who rate themselves as ‘above average’ drivers are four times more likely to be classified as ‘unsafe’ after a 30-minute driving simulation than those who describe their driving ability more modestly (Freund et al., 2005).
These last results present an interesting paradox when it comes to evaluating the benefits and costs of unrealistic self-views. The work of Taylor and others suggest that unrealistically positive self-views are beneficial when it comes to adjusting to or rehabilitating from illness and injury. However, these last findings suggest that unrealistic self-views might prompt people to take risks that lead them to suffer those injuries or setbacks that place them exactly in situations where their overly positive self-views might be called upon to help.
The second central point has to do with the sources of this overestimation. The most intuitive assumption would be that people think so well of themselves simply because they want to. People really do not aspire to accurate views of themselves. They just want pleasant ones. Thus, they are willing to deceive themselves by moulding, massaging, and managing any feedback they receive about themselves until it conforms to their self-images as loveable and capable people.
However, work in my lab suggests that people hold over-inflated self-impressions even when they are sincerely trying to get themselves right. Knowing thyself is an intrinsically difficult task. People often do not have all the information necessary to reach accurate conclusions about the self. Although their self-impressions are often filled with illusion, this is not evidence that people are engaged in willful distortion and delusion.
For example, in many social and intellectual domains, poor performers tend not to know just how badly they perform. They lack insight into their incompetence because their deficits in skill and expertise produce two burdens. The first is the obvious one: lacking skill and knowledge means they make many wrong decisions. The second is not so obvious, but no less important: because they lack the knowledge to recognise what a correct decision looks like, they are not in a position to know that they are making mistakes and that others are choosing more wisely. They don’t know what they don’t know. As a consequence, they believe they are doing just fine – and perhaps even better than most other people.
In our lab, we have found just how little insight poor performers have into their deficiencies. People performing badly on tests of grammar skills, logical reasoning, and their sense of humor typically think they are outperforming a majority of their peers (Kruger & Dunning, 1999). The same goes for students walking out of the classroom after an exam. Those who flunk or nearly flunk think they have outscored most other students in the class. We also see a similar phenomenon in debate teams taking part in a regional collegiate tournament. Poor performing teams estimate that they are winning nearly 60 per cent of their matches when, in fact, they are winning less than a quarter (Ehrlinger et al., 2006).
Our data suggest to us that poor performers are not engaging in willful delusion. Instead, they are providing sincere assessments of how well they think they are doing; they are just not in a position to know how far off these assessments are from the truth. When we remove their gauze of incompetence by training them to become experts, they become much more accurate in their self-assessments and demonstrate a marked willingness to judge themselves harshly (Kruger & Dunning, 1999). Thus, people misjudge their incompetence not because of a lack of honesty with themselves but rather because of a lack of the essential cognitive tools needed to provide correct self-judgments.
Correcting mistaken self-impressions
If gaining accurate self-knowledge is an intrinsically difficult task, what is a person to do? I think there are three steps that people can take to obtain accurate views of self, or at least mitigate their consequences.
The first is to recognise that the road to self-knowledge may run through other people. Our peers might be the best source of the information necessary for accurate self-knowledge. One obvious way to gain self-knowledge is through asking our peers, superiors, and underlings for feedback – and then listening to it. One gentle rule I often tell my students is that if two people independently give them the same piece of negative feedback, they should at least consider the possibility that it might be true.
But even without asking for feedback, others can provide us with self-knowledge in other ways. For example, we can benchmark – compare how other people handle situations with how we handle them. Doing so can often inform us of how we could handle those situations differently – and better – as well as lead us to more accurate views of our skills and expertise.
The second step we can take is to change how we make predictions about ourselves. The usual approach people take to prediction is to spin a story about some outcome (e.g. ‘I’m going to lose two stone in six months’) then to assess its plausibility. A more complete approach would set aside storytelling to look, instead, at data. That is, we could look at whether we have succeeded in similar situations in the past, or look to our peers to see how frequently they succeed in similar circumstances. The key here would be to accept the message the data presents without discounting its relevance if it is uncongenial (Dunning, in press).
The third step would be to just accept that self-knowledge is fraught with error and to adopt ‘cognitive repairs’ to protect oneself from the consequences of those errors (Heath et al., 1998). For example, consider the planning fallacy tendency to overestimate how quickly one can get tasks done. Microsoft knows – painfully – of this tendency, and so when it asks a software developer how quickly he or she can deliver a new piece of software, it automatically applies a cognitive repair, inflating that estimate by 30 per cent (and by 50 per cent if the estimate involves a new operating system).
An unfinished task
Strangers to ourselves? Psychological research suggests that famous aviator Beryl Markham was perhaps right when she said that people can live a lifetime and know other people better than they know themselves. The evidence suggests that self-insight is a precious commodity that people believe they possess to a far greater degree than they really do.
I bet you can all think of many examples of other people who clearly lack adequate self-knowledge. The trick, however, is to be able to tell when we are that person. Finding and ridding ourselves of our own self-misconceptions is a tough business. Thus, we should never be confident that this is a task we have finished for ourselves.
- David Dunning is in the Department of Psychology, Uris Hall, Cornell University, New York. E-mail: [email protected].
Discuss and debate
Are there cultural differences in lack of self-insight and overconfidence?
If most people are overconfident, what produces the underconfidence commonly seen in self and peers?
Why do people make better predictions about other people’s promotions and romances than their own?
Why doesn’t life experience inform the incompetent of their shortcomings? When might it?
What might lead to accurate self-views? Can you think of your own ‘cognitive repairs’?
Have your say on these or other issues this article raises. E-mail ‘Letters’ on [email protected] or contribute to our forum via www.thepsychologist.org.uk.
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. 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.
Buehler, R., Griffin, D. & Ross, M. (2002). Inside the planning fallacy: The causes and consequences of optimistic time predictions. In T. Gilovich, D. Griffin & D. Kahneman (Eds.) Heuristics and biases: The psychology of intuitive judgment (p.251–270). Cambridge, UK: Cambridge University Press.
Cusumano, M.A. & Selby, R.W. (1995). Microsoft secrets. New York: Free Press.
Dunning, D. (2005). Self-insight: Roadblocks and detours on the path to knowing thyself. New York: Psychology Press.
Dunning, D. (in press). Prediction: The inside view. In A. Kruglanski & E.T. Higgins (Eds.) Social psychology: Handbook of basic principles (2nd edition). New York: Guilford.
Dunning, D., Heath, C. & Suls, J.M. (2004). Flawed self-assessment: Implications for health, education, and the workplace. Psychology in the Public Interest, 5, 69–106.
Ehrlinger, J., Johnson, K., Banner, M., Dunning, D. & Kruger, J. (2006). Why the unskilled are unaware: Further explorations of (absent) self-insight among the incompetent. Unpublished manuscript, Stanford University.
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 for behavioral prediction: Enhanced discrimination but exacerbated bias. Personality and Social Psychology Bulletin, 32, 641–655.
Freund, B., Colgrove, L.A., Burke, B.L. & McLeod, R. (2005). Self-rated driving performance among elderly drivers referred for driving evaluation. Accident: Analysis and Prevention, 37, 613–618.
Heath, C., Larrick, R.P. & Klayman, J. (1998). Cognitive repairs: How organizational practices can compensate for individual shortcomings. Research in Organizational Behavior, 20, 1-37.
Jaccard, J., Dodge, T., & Guilamo-Ramos, V. (2005). Metacognition, risk behavior, and risk outcomes: The role of intelligence and perceived knowledge. Health Psychology, 24, 161-170.
Kerby, D.S., Brand, M.W., Johnson, D.L. & Ghouri, F.S. (2005). Self-assessment in the measurement of public health workforce preparedness for bioterrorism or other public health disasters. Public Health Reports, 120, 186–191.
Kruger, J. & Dunning, D. (1999). Unskilled and unaware of it: How difficulties in recognizing one's own incompetence lead to inflated self-assessment. Journal of Personality and Social Psychology, 77, 1121–1134.
Mabe, P.A., III & West, S.G. (1982). Validity of self-evaluation of ability: A review and meta-analysis. Journal of Applied Psychology, 67, 280–286.
MacDonald, T.K. & Ross, M. (1999). Assessing the accuracy of predictions about dating relationships: How and why do lovers’ predictions differ from those made by observers? Personality and Social Psychology Bulletin, 25, 1417–1429.
Malmendier, U. & Tate, G. (2003). Who makes acquisitions? CEO overconfidence and the market’s reaction. Working paper, Stanford University.
Meyer, D., Leventhal, H. & Guttman, M. (1985). Common-sense models of illness: The example of hypertension. Health Psychology, 4, 115–135.
Millis, S.R., Jain, S.S., Eyles, M. et al. (2002). Assessing physicians’ interpersonal skills: Do patients and physicians see eye-to-eye? American Journal of Physical Medicine and Rehabilitation, 81, 946–951.
Risucci, D.A., Torolani, A.J. & Ward, R.J. (1989). Ratings of surgical residents by self, supervisors and peers. Surgical Gynecology and Obstetrics, 169, 519–526.
Taylor, S.E. & Brown, J.D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103, 193–210.
Tracey, J.M., Arroll, B., Richmond, D.E. & Barham, P.M. (1997). The validity of general practitioners’ self assessment of knowledge: Cross sectional study. British Journal of Medicine, 315, 1426–1428.
Weiss, P.M., Koller, C.A., Hess, L.W. & Wasser, T. (2005). How do medical student self-assessments compare to their final clerkship grades? Medical Teacher, 27, 445–449.