Cognitive Biases: Overconfidence

Many pride themselves on their ability to estimate well the probability of this or that situation, or consider themselves logical and analytically precise. Yet at major universities that still offer teaching in philosophy we see subjects such as logic, statistics and even the “critical thinking course”, which is quite popular abroad. This may hint that while we master language at a tender age, logical, that is rational, thinking requires systematic and demanding study. We know that the human mind is not troubled by irrationality alone; often the problem can be that it is better than a rational mind[ref]COSMIDES, Leda and John TOOBY. Better than rational: Evolutionary psychology and the invisible hand. American Economic Review. 84. 1994, : 327-332. Also available at: http://www.cep.ucsb.edu/papers/aer94.pdf, p. 329[/ref]. Both situations arise when our brain is at a loss. It is then forced to resort to mental shortcuts, strategies that we call cognitive biases[ref]In the article I use the English word adapted into Czech; it can be translated, for example, as “illusion” or “distortion”[/ref].

By cognitive biases we mean the known kinds of mental shortcuts that we use when we are forced to decide on the basis of insufficient or invalid data[ref]TVERSKY, Amos and Daniel KAHNEMAN. Judgment under Uncertainty: Heuristics and Biases. Science: New Series. No. 4157. Sep.27,1974, (185): 1124-1131., p.1124[/ref]. We then resort to heuristic methods, which guarantee not the best possible outcome but simplicity and suitability for the problem at hand. These tried and tested patterns of behaviour are useful only in a given context; in other cases they can lead to systematic errors — cognitive biases[ref]GIGERENZER, Gerd. How to Make Cognitive Illusions Disappear: Beyond “Heuristics and Biases”. In: European Review of Social Psychology. 1991, 2(1), pp. 83-115. DOI: 10.1080/14792779143000033. ISSN 1046-3283. Also available at: http://www.tandfonline.com/doi/abs/10.1080/14792779143000033[/ref]. These definitions, taken from Tversky & Kahneman, imply that “human reasoning is irrational, burdened by personal, social or cultural influences”[ref]BERENDSEN, Alina, Simon Jonas HADLICH and Joost VAN AMERSFOORT. Looking at “Linda”: Is the Conjunction Fallacy Really a Fallacy? [online]. [cited 2015-10-18]. Available at: http://bacon.umcs.lublin.pl/~lukasik/wp-content/uploads/2010/12/Looking-at-Linda.pdf[/ref].

That we are prone to such behaviour would point to obvious flaws in the development of our cognition. From the perspective of evolutionary psychology, however, this behaviour can be seen not as “design flaws” but as its characteristic and useful features. Some academics in this subfield of psychology argue that cognitive biases are the result of three evolutionary influences[ref]HASELTON, D., M. G. NETTLE and P.W. ANDREWS. The evolution of cognitive bias. In: BUSS, D. M. (ed.). Handbook of evolutionary psychology. Hoboken: Wiley, 2005, pp. 724-746. Also available at: http://www.sscnet.ucla.edu/comm/haselton/papers/downloads/handbookevpsych.pdf, p. 3[/ref]:

1. useful mental shortcuts that work in most situations (heuristics)
2. distortions that arise when we have to respond to an event the mind was not designed for (artefact)
3. patterns of behaviour that evolved as an evolutionary adaptation to a problem where erroneous action has smaller consequences for the individual’s fitness than logically correct action (error management theory)

Error management theory assumes that many apparently erroneous decisions in human judgement reflect the operation of mechanisms that make us commit many more, and more frequent, relatively “innocent” errors rather than a single fatal one[ref]Ibid.[/ref].

On the basis of EM theory we distinguish biases by three categories: biases concerning judgement under threat, assessment of interpersonal relationships and assessment of oneself[ref]Ibid., p. 10[/ref].

An example of a bias in the last EM category is excessive self-confidence, also known as the overconfidence bias.

The overconfidence bias is, in general, an error of judgement in which an individual is unable to properly assess the limits of their knowledge[ref]SIMON, Mark, Susan M. HOUGHTON and Karl AQUINO. Cognitive biases, risk perception, and venture formation. Journal of Business Venturing. 2000, 15(2): 113-134. DOI: 10.1016/S0883-9026(98)00003-2. ISSN 08839026. Also available at: http://linkinghub.elsevier.com/retrieve/pii/S0883902698000032[/ref].

This definition, however, is not exhaustive; in the literature we find more detailed ones that further divide overconfidence into three distinct kinds of behaviour[ref]MOORE, Don A. and Paul J. HEALY. The Trouble with Overconfidence [online]. [cited 2015-10-18]. Available at: http://repository.cmu.edu/cgi/viewcontent.cgi?article=1340[/ref]:

• Overrating oneself (overestimation)
• Rating oneself above others (overplacement)
• Excessive certainty in one’s judgement (overprecision)

Overestimation is described as excessive confidence in one’s own abilities, performance, level of control or chance of success. Overplacement, the second subtype of the overconfidence bias, occurs when people consider themselves better than others. Overprecision is then understood as too great a certainty that one’s beliefs are correct[ref]Ibid.[/ref].

The literature reports that students often overestimate their performance in exams, doctors overestimate the accuracy of their diagnoses, and people overestimate how much they have things under control, as well as the speed at which they get their work done.

A widely cited example of overplacement is a study in which 93% of American drivers and 69% of Swedish drivers in the sample answered that they felt they were more competent drivers than the average driver of their country[ref]SVENSON, Ola. Are we all less risky and more skillful than our fellow drivers? Acta Psychologica. 1981, 47(2): 143-148. DOI: 10.1016/0001-6918(81)90005-6. ISSN 00016918. Also available at: http://linkinghub.elsevier.com/retrieve/pii/0001691881900056[/ref]. When people are asked to estimate, with 90% certainty, the answers to selected numerical questions (e.g. How long is the river Nile?), their chosen answers turned out to be correct only 50% of the time or less[ref]MOORE, Don A. and Paul J. HEALY. The Trouble with Overconfidence [online]. [cited 2015-10-18]. Available at: http://repository.cmu.edu/cgi/viewcontent.cgi?article=1340, p. 5[/ref].

The practical consequences of this cognitive bias are enormous. Although they do not consider themselves risk-takers, entrepreneurs often start a new business because they lack a sufficient overview of the risk associated with their kind of business; they do not admit to themselves that certain steps necessary for success are beyond their control. In this case, however, the bias can also be beneficial and even necessary for success in the highly competitive environment that entrepreneurs face[ref]SIMON, Mark, Susan M. HOUGHTON and Karl AQUINO. Cognitive biases, risk perception, and venture formation. Journal of Business Venturing. 2000, 15(2): 113-134. DOI: 10.1016/S0883-9026(98)00003-2. ISSN 08839026. Also available at: http://linkinghub.elsevier.com/retrieve/pii/S0883902698000032[/ref]. This is confirmed by Haselton, Nettle et al. when they write that mistaken behaviour and decisions can sometimes lead to heightened ambition and perseverance, whose benefits eventually outweigh the initial negative aspects[ref]HASELTON, D., M. G. NETTLE and P.W. ANDREWS. The evolution of cognitive bias. In: BUSS, D. M. (ed.). Handbook of evolutionary psychology. Hoboken: Wiley, 2005, pp. 724-746. Also available at: http://www.sscnet.ucla.edu/comm/haselton/papers/downloads/handbookevpsych.pdf, p. 22[/ref].

Overconfidence has also been put forward as a possible cause of wars, of risk-taking on the stock market, of the Chernobyl disaster and of the explosion of the Space Shuttle Challenger[ref]MOORE, Don A. and Paul J. HEALY. The Trouble with Overconfidence [online]. [cited 2015-10-18]. Available at: http://repository.cmu.edu/cgi/viewcontent.cgi?article=1340[/ref]. Scott Plous, a professor of social psychology, writes: “No problem in judgement […] has potentially more catastrophic consequences than overconfidence”[ref]PLOUS, Scott. The Psychology of Judgment and Decision Making: (McGraw-Hill Series in Social Psychology). 1st edition. McGraw-Hill, 1993. ISBN 978-0070504776.[/ref].

In all known cognitive biases we can observe correlations with personality characteristics. In the case of the overconfidence bias, research says, for example, that men are more prone to overestimate their abilities, and that empathy and age also correlate positively with overestimation[ref]MOORE, Don A. and Paul J. HEALY. The Trouble with Overconfidence [online]. [cited 2015-10-18]. Available at: http://repository.cmu.edu/cgi/viewcontent.cgi?article=1340, Appendix A[/ref].

Finally, let us mention that Moore and Healy go further and in their work formalise a general theory of confidence, that is, of the way people assess their own performance as well as that of others. They write that after any task people have imperfect information about their own performance and somewhat worse information still about the performance of others. When their own performance is high, people tend to underestimate themselves; they rate the performance of others even somewhat lower, and so they believe they are better than the others. When performance is poor, people’s behaviour and assessments tend to be the opposite: they overestimate their own performance and overestimate that of others even more, which leads them to consider themselves worse than others[ref]Ibid., p. 13[/ref].

Thinking is regarded as the software of our brains. Even the most gifted programmer leaves bugs in their programs, whether out of inattention or on purpose. From evolutionary psychology we learn that these software bugs affect, to a greater or lesser degree, the thinking of every one of us (provided that we consider Darwin’s theory of evolution to be true, although, according to the American presidential candidate Ben Carson, it was inspired by Satan himself). In the history of science we find that too great a certainty of being right, and an inability to give up the paradigm of the moment, have afflicted even the most intelligent members of the human population; let us recall the critics of Copernicus and Galileo, or Albert Einstein, who refused to admit that quantum theory and its “spooky action at a distance” could be true (God doesn’t play dice with the Universe, Einstein said, but quantum entanglement proved him wrong). How can we avoid such errors?

Careful study of cognitive biases, or a printed list of logical fallacies pinned to the noticeboard, is quite a good start. But let us reckon with the fact that a person rarely, if ever, has at their disposal all the data needed to pass a fully informed judgement. That is why intuition, or, if you like, the aesthetic judgement of good form, will play a privileged role even in the life of the most rational person. Without it, we as humanity would still be gathering information about how the world works and would never have dared to take the first step.