The introspection illusion is a cognitive bias in which people wrongly think they have direct insight into the origins of their mental states, while treating others’ introspections as unreliable. In certain situations, this illusion leads people to make confident but false explanations of their own behavior (Causal theories’) or inaccurate predictions of their future mental states.
The illusion has been examined in psychological experiments, and suggested as a basis for biases in how people compare themselves to others. These experiments have been interpreted as suggesting that, rather than offering direct access to the processes underlying mental states, introspection is a process of construction and inference, much as people indirectly infer others’ mental states from their behavior.
When people mistake unreliable introspection for genuine self-knowledge, the result can be an illusion of superiority over other people, for example when each person thinks they are less biased and less conformist than the rest of the group. Even when experimental subjects are provided with reports of other subjects’ introspections, in as detailed a form as possible, they still rate those other introspections as unreliable while treating their own as reliable. Although the hypothesis of an introspection illusion informs some psychological research, the existing evidence is arguably inadequate to decide how reliable introspection is in normal circumstances. Correction for the bias may be possible through education about the bias and its unconscious nature.
The term ‘introspection illusion’ was coined by psychologist Emily Pronin, who describes the illusion as having four components: People give a strong weighting to introspective evidence when assessing themselves; They do not give such a strong weight when assessing others; People disregard their own behavior when assessing themselves (but not others); and Own introspections are more highly weighted than others (it is not just that people lack access to each other’s introspections — they regard only their own as reliable).
A 1977 paper by psychologists Richard Nisbett and Timothy D. Wilson challenged the directness and reliability of introspection, thereby becoming one of the most cited papers in the science of consciousness. Nisbett and Wilson reported on experiments in which subjects verbally explained why they had a particular preference, or how they arrived at a particular idea. On the basis of these studies and existing attribution research, they concluded that reports on mental processes are confabulated (fabricated, distorted, or misinterpreted). They wrote that subjects had, ‘little or no introspective access to higher order cognitive processes.’ They distinguished between mental contents (such as feelings) and mental processes, arguing that while introspection gives us access to contents, processes remain hidden.
Although some other experimental work followed from the Nisbett and Wilson paper, difficulties with testing the hypothesis of introspective access meant that research on the topic generally stagnated. A ten-year-anniversary review of the paper raised several objections, questioning the idea of ‘process’ they had used and arguing that unambiguous tests of introspective access are hard to achieve. Updating the theory in 2002, Wilson admitted that the 1977 claims had been too far-reaching. He instead relied on the theory that the adaptive unconscious does much of the moment-to-moment work of perception and behavior. When people are asked to report on their mental processes, they cannot access this unconscious activity. However, rather than acknowledge their lack of insight, they confabulate a plausible explanation, and ‘seem’ to be ‘unaware of their unawareness.’
The idea that people can be mistaken about their inner functioning is one applied by eliminative materialists. These philosophers suggest that concepts, including ‘belief’ or ‘pain’ will turn out to be quite different than commonly expected as science advances. The faulty guesses that people make to try and explain their thought processes have been called ‘Causal theories.’ The causal theories provided after an action will often serve only to justify the person’s behavior in order to relieve Cognitive dissonance (discomfort experienced when holding contradictory thoughts). That is, a person may not have noticed the real reasons for their behavior, even when trying to provide explanations. The result is an explanation that mostly just makes themselves feel better. An example might be a man who discriminates against homosexuals because he is embarrassed that he himself is attracted to other men. He may not admit this to himself – instead claiming his prejudice is because he believes homosexuals are ‘not natural’ in some vague sense.
A study conducted by philosopher Eric Schwitzgebel and psychologist Russell T. Hurlburt was set up to measure the extent of introspective accuracy by gathering introspective reports from a single individual who was given the pseudonym ‘Melanie.’ Melanie was given a beeper which sounded at random moments, and when it did she had to note what she was currently feeling and thinking. After analyzing the reports the authors had mixed views about the results, the correct interpretation of Melanie’s claims and her introspective accuracy. Even after long discussion the two authors disagreed with each other in the closing remarks, Schwitzgebel being pessimistic and Hurlburt optimistic about the reliability of introspection.
Nisbett and Wilson conjectured about several factors that they found to contribute to the accuracy of introspective self-reports on cognition. Availability (stimuli that are highly salient, either due to recency or being very memorable, are more likely to be recalled and considered for the cause of a response); Plausibility (whether a person finds a stimulus to be a sufficiently likely cause for an effect determines the influence it has on his or her reporting of the stimulus); Removal in time (the greater the distance in time since the occurrence of an event, the less available and more difficult to accurately recall it is); Mechanics of judgment (people do not recognize the influence that judgment factors have on them, leading to inaccuracies in self-reporting); Context (focusing on the context of an object distracts from evaluation of that object and can lead people to falsely believe that their thoughts about the object are represented by the context); Nonevents (the absence of an occurrence is naturally less salient and available than an occurrence itself, leading nonevents to have little influence on reports); Nonverbal behavior (while people receive a large amount of information about others via nonverbal cues, the verbal nature of relaying information and the difficulty of translating nonverbal behavior into verbal form lead to a lower reporting frequency of this behavior); and Discrepancy between the magnitudes of cause and effect (because it seems natural to assume that a certain size cause will lead to a similarly-sized effect, connections between causes and effects of different magnitudes are not often drawn).
Several hypotheses to explain people’s unawareness of their inaccuracies in introspection were provided by Nisbett and Wilson: Confusion between content and process (people are usually unable to access the exact process by which they arrived at a conclusion, but can recall an intermediate step prior to the result); Knowledge of prior idiosyncratic reactions to a stimulus (an individual’s belief that he or she reacts in an abnormal manner to a stimulus, which would be unpredictable from the standpoint of an outside observer, seems to support true introspective ability); Differences in causal theories between subcultures (the inherent differences between discrete subcultures necessitates that they have some differing causal theories for any one stimulus); Attentional and intentional knowledge (an individual may consciously know that he or she was not paying attention to a certain stimulus or did not have a certain intent); Inadequate feedback (by nature, introspection is difficult to be disconfirmed in everyday life, where there are no tests of it and others tend not to question one’s introspections – moreover, when a person’s causal theory of reasoning is seemingly disconfirmed, it is easy for him or her to produce alternative reasons for why the evidence is actually not disconfirmatory at all); and Motivational reasons (considering one’s own ability to understand his or her reasoning as being equivalent to an outsider’s is intimidating and a threat to the ego and sense of control).
Inspired by the Nisbett and Wilson paper, Petter Johansson and colleagues investigated subjects’ insight into their own preferences using a new technique. Subjects saw two photographs of people and were asked which they found more attractive. They were given a closer look at their ‘chosen’ photograph and asked to verbally explain their choice. However, using sleight of hand, the experimenter had slipped them the other photograph rather than the one they had chosen. A majority of subjects failed to notice that the picture they were looking at did not match the one they had chosen just seconds before. Many subjects confabulated explanations of their preference. For example, a man might say ‘I preferred this one because I prefer blondes’ when he had in fact pointed to the dark-haired woman, but had been handed a blonde.
The large proportion of subjects who were taken in by the deception contrasts with the 84% who, in post-test interviews, said that hypothetically they would have detected a switch if it had been made in front of them. The researchers coined the term ‘choice blindness’ for this failure to detect a mismatch. It is not clear, however, the extent to which these findings apply to real-life experience when there is time to reflect. As Prof. Kaszniak points out: ‘although a priori theories are an important component of people’s causal explanations, they are not the sole influence, as originally hypothesized by Nisbett & Wilson. Actors also have privileged information access that includes some degree of introspective access to pertinent causal stimuli and thought processes, as well as better access (than observers) to stimulus-response covariation data about their own behavior.’
Studies that ask participants to introspect upon their reasoning (for liking, choosing, or believing something, etc.) tend to see a subsequent decrease in correspondence between attitude and behavior in the participants. For example, in a study by Wilson et al. participants rated their interest in puzzles that they had been given. Prior to the ratings, one group had been instructed to contemplate and write down their reasons for liking or disliking the puzzles, while the control group was given no such task. The amount of time participants spent playing with each puzzle was then recorded. The correlation between the puzzle ratings and time spent playing with each was much smaller for the introspection group than the control group.
A subsequent study was performed to show the generalizability of these results to more ‘realistic’ circumstances. In this study, participants were all involved in a steady romantic relationship. All were asked to rate how well-adjusted their relationship was. One group was asked to list all of the reasons behind their feelings for their partner, while the control group did not do so. Six months later, the experimenters followed up with participants to check if they were still in the same relationship. Those who had been asked to introspect showed much less attitude-behavior consistency based upon correlations between the relationship ratings and whether they were still dating their partners (i.e., introspection was not predictive).
The authors theorize that these effects are due to participants changing their attitudes, when confronted with a need for justification, without changing their corresponding behaviors. The authors hypothesize that this attitude shift is the result of a combination of things: a desire to avoid feeling foolish for simply not knowing why one feels a certain way; a tendency to make justifications based upon cognitive reasons, despite the large influence of emotion; ignorance of mental biases (e.g., halo effects); and self-persuasion that the reasons one has come up with must be representative with his or her attitude. In effect, people attempt to supply a ‘good story’ to explain their reasoning, which often leads to convincing themselves that they actually hold a different belief. In studies wherein participants chose an item to keep, their subsequent reports of satisfaction with the item decreased, suggesting that their attitude changes were temporary, returning to the original attitude over time.
The bias blind spot is an established phenomenon that people rate themselves as less susceptible to bias than their peer group. Emily Pronin and Matthew Kugler argue that this phenomenon is due to the introspection illusion. In their experiments, subjects had to make judgments about themselves and about other subjects. They displayed standard biases, for example rating themselves above the others on desirable qualities (demonstrating illusory superiority). The experimenters explained cognitive bias, and asked the subjects how it might have affected their judgment. The subjects rated themselves as less susceptible to bias than others in the experiment (confirming the bias blind spot). When they had to explain their judgments, they used different strategies for assessing their own and others’ bias.
Pronin and Kugler’s interpretation is that when people decide whether someone else is biased, they use overt behavior. On the other hand, when assessing whether or not they themselves are biased, people look inward, searching their own thoughts and feelings for biased motives. Since biases operate unconsciously, these introspections are not informative, but people wrongly treat them as reliable indication that they themselves, unlike other people, are immune to bias. Pronin and Kugler tried to give their subjects access to others’ introspections. To do this, they made audio recordings of subjects who had been told to say whatever came into their heads as they decided whether their answer to a previous question might have been affected by bias. Although subjects persuaded themselves they were unlikely to be biased, their introspective reports did not sway the assessments of observers.
Another series of studies by Pronin and colleagues examined perceptions of conformity. Subjects reported being more immune to social conformity than their peers. In effect, they saw themselves as being ‘alone in a crowd of sheep.’ The introspection illusion appeared to contribute to this effect. When deciding whether others respond to social influence, subjects mainly looked at their behavior, for example explaining other student’s political opinions in terms of following the group. When assessing their own conformity, subjects treat their own introspections as reliable. In their own minds, they found no motive to conform, and so decided that they had not been influenced.
Psychologist Daniel Wegner has argued that an introspection illusion contributes to belief in paranormal phenomena such as psychokinesis. He observes that in everyday experience, intention (such as wanting to turn on a light) is followed by action (such as flicking a light switch) in a reliable way, but the processes connecting the two are not consciously accessible. Hence though subjects may feel that they directly introspect their own free will, the experience of control is actually inferred from relations between the thought and the action. This theory, called ‘apparent mental causation,’ acknowledges the influence of David Hume’s view of the mind. This process for detecting when one is responsible for an action is not totally reliable, and when it goes wrong there can be an illusion of control. This could happen when an external event follows, and is congruent with, a thought in someone’s mind, without an actual causal link.
As evidence, Wegner cites a series of experiments on magical thinking in which subjects were induced to think they had influenced external events. In one experiment, subjects watched a basketball player taking a series of free throws. When they were instructed to visualize him making his shots, they felt that they had contributed to his success. If the introspection illusion contributes to the subjective feeling of free will, then it follows that people will more readily attribute free will to themselves rather than others. This prediction has been confirmed by three of Pronin and Kugler’s experiments. When college students were asked about personal decisions in their own and their roommate’s lives, they regarded their own choices as less predictable. Staff at a restaurant described their co-workers’ lives as more determined (having fewer future possibilities) than their own lives. When weighing up the influence of different factors on behavior, students gave desires and intentions the strongest weight for their own behavior, but rated personality traits as most predictive of other people.
The Daily Omnivore 