2. Errors and biases Flashcards
Explain the fundamental attribution error.
When people have to attribute causality to an event they have witnessed, they tend to attribute cause to the person involved rather than the situation factors that may have shaped how the person acted.
The FAE is a general tendency that is acquired through socialisation into Western cultures for observers to overestimate
the importance of personal/
dispositional factors relative to environmental influences.
What did Ross suggest regarding the fundamental attribution error?
That everybody is a ‘naive psychologist trying to understand their own behaviour and that of other people. Naive psychologists are particularly prone to making the FAE.
How might individuals differ in how they attribute the cause of behaviour?
Individuals performing the behaviour may attribute cause of their behaviour differently to those observing. Actors tend to attribute behaviour to situational factors and observers tend to attribute it to dispositional factors.
The reason actors and observers differ in attributing causality is due to tehir kind of attentional focus. If performing a behaviour, your attention is focused outwards- we see the world around us but not ourselves and so we are more aware of situational factors. However, observers are looking at the people doing the behaviours and so are more likely to pay attention to dispositional factors.
Explain how the FAE differs between cultures.
Miller proposed that the FAE only applies to western cultures. He conducted a cross cultural study to investigate whether the FAE would hold across cultures using 7 participants form India and 60 from the US.
Western culture is individualistic, individuals are self-defined by internal attributes, and encourages separation of the self from context. Indian culture is collectivistic, individuals are self-defined by social relationships, and encourages integration of the self into social contexts.
If the FAE is culture bound, we would expect those from the USA to be more prone to it and for these differences to emerge with age as we are socialised into the culture. Both children (aged 8, 11, or 15) and adults were studied.
Findings: Among adults, participants from the US attributed greater causality ot actors predispositions for both pro and anti social behaviours. The results of 15 year olds resembled those of adults, however there was no difference between the findings of 8 & 11 year olds from the different cultures.
There are important cultural differences in susceptibility to FAE; observers from a culture that is prototypically individualistic are more likely to attribute causality to disposition. These cultural differences are learned from socialisation into ones culture and occurs between 11 and 15.
Explain the logical arguments against the FAE.`
Is the FAE an error?: we can’t know the true cause of behaviour based on just one instance but we usually don’t get to observe something multiple times. There is no objective criteria for identifying a causal agent; the FAE may be a bias but not an error because observers are not making objectively incorrect causal attributions. The FAE has been recently relabelled the correspondence bias.
The logical coherence of the correspondence bias has been questioned:
Every internal cause has a matched external one and vice=versa. Sabini et al. argue that behaviour always arises from an interaction between an actor and the situation and that the two cannot be separated as the actor has to act within the situation. The better way to think about causality of any action is like 2 magnets: both poles cause the attraction because they both have causal influences.
Why does correspondence bias happen?
Gilbert et al proposed a model of the correspondence bias to explain why it happens. it involves 3 sequential stages in making attributions.
- Categorisation: we notice an action and seek to categorise it into something meaningful
- Characterisation: what does the action imply about the actor?
- Correction: what situational factors might have caused this action?
The categorisation and characterisation process are undertaken automatically without conscious thought but the correction process requires conscious cognitive effort. We often lack the capability or motivation required to correct our misconceptions.
Explain the test Gilbert et al conducted to study their explanation.
This test involved imposing a cognitive load on participants. Participants should be unable to correct their initial characterisations when under cognitive load as they do not have the capacity to perform the correction process- those under cognitive load should be more likely to attribute causality to the actor.
They partially repeated the Castro study, however to be more up to date they used abortion instead.
Participants were asked to listen to either an anti or pro abortion speech and were told that the writer had been instructed to write the speech for that side of the argument. They were then asked to rate the writers’ true attitudes towards abortion on a scale of 1 to 13, with one being strong anti abortion and 13 being strong pro abortion.
50 percent of participants were given a cognitive load by being asked to prepare their own speech on abortion- they were using their mental resources on their own speech and so don’t have these resources available to make accurate attributions.
Research question: would participants under a greater cognitive load show more susceptibility to correspondence bias?
Results: Participants without cognitive load were more likely to infer pro-abortion attitudes among those who wrote a pro essay than an anti abortion essay and vice versa. Participants under a cognitive load found a larger difference, rating those who had written pro-essays to be more pro than the other group rated them This shows that participants under a cognitive load were more susceptible to correspondence bias.
Explain using cognitive load in studies.
This test involved imposing a cognitive load on participants. This involves giving participants a mentally effortful task so we can see how well they perform another task.
This works on the bases that we have a limited amount of mental resources available for any one task at any one time. When we don’t have sufficient mental resources, we fall back on automatic processes which allow us to do tasks while using limited mental resources- this reveals how people perform when on autopilot: reveals bias.
Explain bounded rationality.
Traditional economic theories of human reasoning assume:
- We have a well organised system of stable preferences
- We always make the decision that will yield the outcome we most prefer based on consulting our knowledge of relevant aspects of our environment
Simon argues for a more realistic appraisal of human rationality:
Bounded rationality is the idea that we try to be rational but we cannot be fully rational as our capacity for being rational is limited by availability of information, time and mental capacity/resources. We are as rational as we can be.
Explain confirmation bias.
Snyder and Swann: designed a study to look at how people test hypotheses in social interactions.
Participants had to interview a fellow student to find out more about them. Half the participants were told to identify whether the student was introverted, shy, quiet or timid while the latter half were told to identify whether their partner was extroverted, outgoing, sociable and energetic. Half of the participants from each condition were told that this information had been drawn from a personality test taken by the student last week- they were led to believe that there was a high certainty the hypothesis was true. The other half were simply told to find out whether the term introverted or extroverted was fitting- they were led to believe there was less certainty about the hypothesis. Participants were then asked to select questions they wanted to ask from a list (some more relevant to introversion, some more relevant to extroversion, some were neutral).
It was found that those testing whether the participant was an extrovert chose a greater number of extroversion questions, and vice versa for those testing for introversion.
There was no difference in questions chosen based on hypothesis certainty.
These results tell us that when people test hypotheses about other people, they tend to be more biased towards seeking out information that is consistent with their original hypothesis. The effect held in subsequent studies, even when participants were incentivised with money to be accurate.
Explain a following study to that of Snyder and Swann.
Participants actually asked their questions rather than just selecting them, and a second participant listened in on the interview and had to rate the introversion or extroversion of the participant. Participants were told to be open, honest and consistent in their answers.
It was found that where the participant interviewer asked more questions about extroversion, the observing participant rated the participant being interviewed as being more extroverted and vice versa for introversion.
These results show the potential for the self-fulfilling prophecy to occur. A possible consequence of confirmation bias when questioning people is that through the questioning, we may conclude judgements that are not correct.
Explain the studies of confirmation bias in the interrogation room.
Participants were led to believe that participants were either guilty or innocent of a mock theft and were then allowed to choose which questions they would ask the suspect and which interrogation techniques they would use.
Those whose initial hypothesis was that the suspect was guilty chose more questions that presumed guilt and used harder interrogation techniques to try and extract a confession.
In the second experiment neutral participants were led to believe that suspects were guilty or innocent and listened to tapes of interviews with suspects. Where the participants believed the suspect to be guilty, they interpreted their response as being more defensive and rated the likelihood for them to be guilty higher.
This suggests that the initial assessment led to the decision made about the suspect being biased.
Evaluate the studies of confirmation bias.
A weakness of Snyder and Swann’s study is the methodology- participants were instructed to confirm the hypothesis, not test it.
Other studies have shown that instructing participants to test whether a hypothesis is false instead can reduce or eliminate the confirmation bias.
The police role might have encouraged participants to actively seek out disconfirmatory information as they were taking the position of someone whose job is to seek out and punish violations of the rule.
However, many other studies have also found evidence for confirmation bias. The limitation doesn’t mean that the study is completely invalid but perhaps just that the methods Snyder and Swann used amplified the findings.
Explain some of the explanations for why confirmation bias occurs.
Information processing explanation:
- Has the confirmation bias arisen from an inability to process too much information at once? Focusing on a state of affairs where the hypothesis is true means we might struggle to also focus on a state of affairs where the hypothesis is false and so confirmation could arise from our schemas.
Pragmatism:
- If our hypothesis is true, we should be able to find both confirmatory information and that which goes against it. It has been suggested that if we are unable to find confirmatory information then our hypothesis is more wrong than if we were able to find disconfirmatory and confirmatory information. We don’t want to make obvious mistakes and we see a failure to find confirmation as a more basic mistake than failure to find recognised disconfirmation.
Educational effect:
In Western culture, our education system instills into us a notion that we must justify what we believe and provide evidence for how to arrive to our conclusions which biases us more to seeking out hypothesis consistent information than hypothesis inconsistent information.
Explain confirmation bias in a high-speed sport scenario.
In some sports you cannot perform without anticipating.
Skilled performers use information available early on to help anticipate what’s going to happen. This might be medical history of a patient, tendencies, previous behaviour…this enables performers to act faster in situations where the information is actually correct.
If this information is wrong (incongruent), however, we found that skilled performers were found to be less accurate than non-skilled performers at anticipation. Less skilled performers don’t have the level of understanding to use early information in the same way and so are relying on visual information- their accuracy doesn’t change from congruent to incongruent.
This shows that there may be confirmation bias occurring, where players make conclusions based on early information, but this only works well if the early information is congruent. They might not be updating their anticipations seeking out information that might prove it wrong.
Due to the short time period they have, they may only have time to look for confirmatory information but not information that would prove the original prediction wrong.
