Flawed Self-Assessment [Dunning, Heath and Suls]
It is often the case that in our push for change, we tend to go overboard and “throw the baby with the bathwater”. One of the hottest trends in People Operations, the overhaul (and sometime the complete removal) of any sort of performance assessment and management systems is not immune to that common pitfall.
The extreme argument posits that since we are all fully-grown adults, we have a clear understanding of our strengths and weaknesses, and therefore completely capable of driving our own professional development on our own. The system, or the organization, should just take its cue from us, and support us in pursuing whichever direction we see fit.
In this 40-page meta-analysis David Dunning, Chip Heath and Jerry Suls clearly lay out the challenges that prevent us from accurately self-assess our abilities, in making decisions about our health, our education and our work.
If you’re not up to reading the 40-pager, here’s an outline of the most relevant sections of their paper to the reader of this publication, copied almost verbatim from the paper with minor editorial changes to ease readability.
Empirical evidence exists for some major flaws of in people ability to self-assessment.
Overall, there is a low correlation between perception and reality:
- In the workplace, the correlation between how people expect to perform and how they actually perform hovers around .20 for complex tasks
- Some domains produce higher correlations than others. In athletics, where feedback tends to be constant, immediate, and objective, the typical correlation was .47. In the realm of complex social skills, where feedback might be occasional and is often delayed and ambiguous, it tended to be much lower (e.g., .04 for managerial competence and .17 for interpersonal skills)
- Peer ratings of leadership, rather than self-ratings, predict which naval officers will be recommended for early promotion.
Specifically, unrealistic optimism is pervasive:
- Above-average effects — People, on average, tend to believe themselves to be above average — a view that violates the simple tenets of Mathematics.
- Overestimation of the likelihood of desirable events — People overestimate their ability to bring about personally desirable events. People also overestimate the likelihood that their own future actions will be socially desirable, even though their predictions regarding their peers’ behavior turn out to be more accurate.
- Underestimation of task-completion times — People consistently overestimate how easily they can complete tasks (as measured by time or money), a phenomenon known as the planning fallacy.
- Overconfidence in judgment and prediction — People place too much confidence in the insightfulness of their judgments, overestimating the chances that their decisions about the present are sound and that their predictions about the future will prove correct. This phenomenon is known as the overconfidence effect.
The psychological mechanisms that underlie flawed self-assessments can be grouped under two major themes. The first theme is that people typically do not possess all the information required to reach perfectly accurate self-assessments. The second theme is that even when people do have valuable information that would guide them toward appropriate self-evaluations, they often neglect it or give it too little weight; thus, they make potentially avoidable errors.
Explanations for the above-average effect:
- Information deficits — The Double Curse of Incompetence. People often do not have the knowledge and expertise necessary to assess their competence adequately.
- Unknown errors of omission — Although people trying to solve a problem may find it easy to consciously critique the solutions they generate, by definition they are not aware of solutions they could generate but miss, that is, their errors of omission.
- Uncertain lessons from feedback — People receive incomplete feedback about their actions, which can lead them to harbor inflated views about the wisdom of their actions.
- The ill-defined nature of competence — Perhaps the most fundamental reason for people to have incomplete knowledge of their competence is that in many domains, what it takes to succeed is hard to define.
- Information neglect — People misjudge themselves relative to others because they ignore crucial information, and this neglect can produce the above-average effect and, on occasion, its direct opposite.
- Exclusive focus on the self with neglect of others — People’s comparative judgments often involve very little comparison. When evaluating their skill vis-a`-vis their peers’, people are egocentric, thinking primarily of their own behaviors and attributes and ignoring those of others.
- Controllability and privacy of traits — People think of themselves as superior to their peers when thinking about traits that are construed as controllable,but not so much when thinking about uncontrollable traits. People consider themselves more cooperative and self-disciplined than others (all controllable qualities), but not necessarily more creative or lively. People tend to believe that they possess traits more than their peers to the extent that those traits tend to be expressed internally. For example, people tend to think they are more self-conscious, self-critical, and choosy than their peers, but not that they are more aggressive, poised, or wordy, traits that are more external in their expression. The neglect of the internal lives of others can result in a phenomenon known as pluralistic ignorance, in which people believe they uniquely possess a deviant opinion, whether desirable or not, when in reality most people in the community privately share the same opinion.
Explanations for overly optimistic predictions of events:
- Information deficits — Unknown situational details: People often make overly confident predictions, typically optimistic ones but not necessarily so, because they fail to correct for the fact that the details of future situations are often unknown or unpredictable, even though those details may matter. As well as imperfect understanding of emotion, visceral drives, and their consequences: Situational features are inaccessible in other ways. People often have difficulty predicting how they will respond to situations that have significant emotional or visceral components.
- Neglect of alternative scenarios — People mispredict future events because they neglect important information that they have in hand. When they spin scenarios about how they will behave in the future, they tend to dwell on positive scenarios and fail to take into account worst-case scenarios that they could easily generate.
- Neglect of concrete detail -People also base their predictions about events in the distant future on abstract, higher-level features of a situation and give short shrift to more concrete, lowlevel features that can have a significant impact on behavior.
- Neglect of background circumstances — People mispredict because their imagined scenarios concentrate too much on the behavior in question and not about seemingly irrelevant swirls and eddies of everyday life that are not conceptually related to the behavior but that may still interfere with their capacity to perform that behavior.
- Neglect of the lessons of experience — People ignore their previous experience because they generally take an ‘‘inside view’’ rather than an ‘‘outside view’’ when predicting how quickly they will complete tasks . When people take the inside view, they consider the unique features of the task at hand and imagine a series of steps that will lead them from their starting point to a solution. As a consequence, they focus on their abilities and resources, perhaps envisioning obstacles and thinking about how they will overcome them. In contrast, when people take an outside view, they dismiss this scenario building focusing on the situation at hand and instead pursue a more data-driven strategy in which they just tally the final outcomes from situations they know of that are similar to the one they now face.
Implications to the workplace:
Accuracy of self-knowledge in organizations
The largest surprises generated by lack of self-knowledge may be those that are produced when self-evaluations are not echoed by supervisors who set raises and hand out promotions. If employees overrate their own performance, it is difficult to imagine how people could not wind up disappointed at least some of the time.
- Zenger (1992) studied several hundred engineers at two high-tech companies and found that 32% of the engineers in one company and 42% in the other rated their own performance in the top 5% of all engineers.
- Typically, the views of other people — subordinates, peers, and superiors — agree with each other more often than with self- views. Although the ratings of peers and supervisors agreed pretty well ( r = .62), their correlation with people’s own ratings of their job performance was lower ( r = .35 for supervisors’ ratings, r= .36 for peers’ ratings)
Challenges in providing employees with feedback
Actual feedback systems can become ineffective for the following reasons:
- Feedback is infrequent
- Feedback is threatening
- Feedback is sugarcoated
- Feedback is given too late
The burden of feedback systems falls especially hard on front-line managers, who have to give feedback that disappoints or angers employees who assume they are above average and cannot see why their managers do not agree. Because of these problems with giving feedback, many organizations evolve to an equilibrium pattern of feedback that may be suboptimal for organizational performance, but more viable from the standpoint of interpersonal relationships.
How, then, can organizations recognize and develop their best employees without undermining the motivation of the bulk of employees in the middle? One way might be to give high and low evaluations to a few exceptionally good and bad performers, respectively, and to give the bulk of the employees ambiguously positive evaluations. When any particular employee is singled out for special positive recognition, this makes many other employees — all those who correctly or incorrectly feel they are performing better — feel that their efforts have not been equitably rewarded. Organizations with positively skewed evaluation systems may get away with singling out a few really high performers — those who peers agree are clearly deserving — if they avoid tougher distinctions among people in the middle of the distribution, where quality distinctions may be more contested. Such evaluation systems cleverly allow organizations to select the best (and get rid of the worst) with the fewest ruffled feathers, but the problem is that these systems do not provide people in the middle of the distribution with an evaluation that may signal them to improve their performance. The overall results of such systems may depend on whether organizations can separate feedback from evaluation and provide feedback that encourages people in the middle to improve even though they are already being evaluated as ‘‘above average.’’
Cognitive repairs for mistaken self-judgment
For people at the bottom of a hierarchical organization, lack of self-knowledge may hinder career or personal success. Organizations often develop or evolve reminders, routines, and procedures that help mitigate problems that could be caused by employees who lack awareness of their own abilities:
- Adding safety factors and buffer time
- Forcing people to pay attention to the environment (“Don’t confuse brains and a bull market”)
- Providing feedback from other people: Disney’s “Gong Show”, Pentagon’s “Murder Board”
Some common themes run through the literature on improving the accuracy of self-judgment. One theme that emerges from our review is that the road to self-accuracy may involve information from or about other people. Another theme, coming from the organizational literature, is that cognitive repairs can be applied to the kinds of self-judgments that are often made with error, thus sparing individuals and their organizations the costs associated with faulty self-assessment.