by Paul Seager
‘One of the greatest and most remarkable misunderstandings we have about people, one that gives rise to many other inferential failings, is the belief that behavior is usually best regarded as reflecting personality traits or other internal attributes.’
(Choi, Nisbett & Norenzayan,1999, p. 47)
Some research suggests that we make such judgements on the basis of very little information indeed. For example, one study asked individuals to talk to other individuals on the phone, and then describe the people that they had talked to. The results showed that the participants made judgements about the stranger’s personality trait just based on hearing their voice. Imagine that you are talking to a person on the phone for the very first time (whether it be a customer for your business, or a person on a sales help line) – it is likely that by the time you put the phone down, you will have started to ‘form a picture’ (make attributions) of that person.
Heider laid many of the foundations of modern attribution theory on which more formal theories of attribution have sought to build. We will now look in more depth at two such theories.
Jones and Davis (1965) theory of correspondent inference
One of the first theories to be proposed was by Jones and Davis (1965) who wanted to show how people make a connection between an intentional overt action and an internal trait; how an individual infers that a deliberate action by another person corresponds to an underlying disposition. For example, if a lecturer is always shouting at students in seminars, or a co-worker is always shouting at his colleagues (we assume that the lecturer or co-worker chooses to perform this behaviour), then we are likely to infer that they have an ‘angry’ disposition. Again we do this because we like things to be neat and tidy, and as simple as possible – it’s easier to ‘pigeonhole’ somebody (‘He is an angry person’) than to deal with the reality that things might be a bit more complex. It’s easier to deal with someone if we ‘know’ what they are like. It gives us the control over our world to which Heider was referring.
Key idea: Correspondent inference theory
The idea that an overt action (behaviour) of another person is causally related to an internal disposition that they possess.
According to Jones and Davis’ theory, we attempt to link an action to a specific internal cause, and to make these correspondent inferences, we use a number of sources of information:
1 IS THE BEHAVIOUR FREELY CHOSEN?
If a person’s behaviour appears to be ‘freely chosen’, then it is more likely to correspond to an underlying disposition. For example, if we see Celia in the park playing football in the rain, we are more likely to judge that she likes football, as we assume that very few people would choose to play in the rain without such a liking. However, if we judge that a person’s actions are due to an external source, then we are less likely to infer that the behaviour corresponds to a disposition. For example, if we see Celia walking a dog in the rain, but we know that her Mum insists that she take her turn to walk the dog, we wouldn’t necessarily infer that Celia is a dog-lover, or someone who likes to walk pets in the rain. Of course, we won’t always know whether an action is freely chosen or not.
2 DOES THE BEHAVIOUR LEAD TO A SPECIFIC EFFECT?
If a person’s behaviour has a very unique outcome (a non-common effect), as opposed to a range of possible consequences, then we are more likely to infer an internal disposition for that behaviour. For example, if a person walks into a bank with a gun drawn, then there are likely to be very few outcomes from that behaviour (and most involving a violent outcome), and therefore we are likely to infer that the person has a violent, or criminal, disposition. However, if a person walks into a supermarket with a ten-pound note in their hand, there are a whole range of reasons for this behaviour, and therefore we are less likely to make an internal attribution.
3 IS THE BEHAVIOUR SOCIALLY DESIRABLE OR NOT SOCIALLY DESIRABLE?
If a person’s behaviour is deemed to be socially desirable, then it tells us little about their internal disposition as their actions are likely to be driven by societal norms. For example, if we see a person going into a polling station to vote on general election day, then we wouldn’t necessarily draw any inferences about their behaviour as voting is generally seen as the right thing that we should all do. However, if we saw the same person go into the polling station and deliberately ‘spoil’ their ballot paper, then as this behaviour is fairly unusual and not necessarily socially desirable, we are more likely to attribute their behaviour to an internal disposition, for example perhaps they have ‘anarchic’ tendencies.
‘Given an inference that assigns an attribute to account for an act, the correspondence of that inference increases as the judged value of that attribute departs from the judge’s conception of the average person’s standing.’
(Jones & Harris,1967, p. 2)
Other sources of information have also been put forward as being influential in leading to people making correspondent inferences, and these include behaviour that is high in hedonic relevance (the person’s behaviour has important consequences for us as an individual) and behaviour that seems to be ‘out-of-role’ (the person’s behaviour is not typical for the role that we perceive them in e.g. nurse, policeman, banker).
Whilst a number of studies have provided evidence to support aspects of this theory (see Case study below), it has tended to wane in popularity due to a number of limitations. For example, it suggests that inferences are made on the assumption of intentionality (i.e. the behaviour of the individual being observed was carried out deliberately), but there are cases where unintentional behaviours (e.g. carelessness) can lead to definite inferences. Also, the theory tends to concentrate a little too much on internal dispositions being the cause of behaviour, as opposed to external causes. The next theory of attribution specifically aims to address such problems.
Case study: The attribution of attitudes
Jones & Harris (1967) conducted an experiment to ‘attempt to determine if people can make valid judgements of another’s personality and attitudes on the basis of very limited information’ (p. 4). Participants were asked to read a 200-word essay, allegedly written by a student regarding their views on Fidel Castro (remember this was the 1960s when Castro’s Cuba was seen as undesirable by much of the western world). There were four conditions in the experiment – the essay was either ‘for’ or ‘against’ Castro, and the writer was either given a free choice of which view to take, or forced to write a specific viewpoint. Having read the essay, participants were asked to make a number of judgements about the writer’s personality and dispositions. The crucial issue was when participants were asked a number of questions about the writer’s attitude to Castro.
In both conditions where the essay was ‘anti-Castro’, participants judged the writer as having negative attitudes towards Castro. In the pro-Castro condition, where the writer was given a choice of which view to take, the participants judged the writer as holding ‘pro-Castro’ attitudes, in line with correspondent inference theory. However, surprisingly, in the pro-Castro condition, where the writer was forced to write the pro-Castro essay, participants also judged the writer as holding pro-Castro attitudes, thus showing that participants were making attributions based on internal dispositions, even though there were external ‘situational’ causes for the writer’s behaviour.
Kelley’s (1967) co-variation model
According to Kelley’s co-variation model, in order to attribute a cause to an observed behaviour, people look for which factors co-vary most closely with an observed behaviour; they look to see if the suspected cause of the behaviour is always present when the behaviour is present and is absent when the behaviour is absent. Whilst behaviour could have multiple causes, the cause which co-varies (correlates) the closest with the behaviour is more likely to be attributed as the reason for the behaviour. We are acting like rational thinkers and making objective observations about behaviour in order to arrive at a conclusion for why the behaviour occurre
d (think back to the idea of Heider’s naïve scientists).
Key idea: Co-variation model
An attempt to show how, using three sources of information, people attribute the cause of an individual’s target behaviour with the factor that co-varies most closely with the behaviour.
The co-variation model suggests that, in order to determine whether the cause of the behaviour should be attributed to internal (dispositions) or external (situational) factors, people use three types of information:
1 Consistency information: Does the behaviour always occur in the presence of the potential cause (high consistency) or only sometimes (low consistency)?
2 Distinctiveness information: Is the behaviour only seen when the potential cause is present (high distinctiveness) or is it seen at any other times too (low distinctiveness)?
3 Consensus information: Does everyone carry out the behaviour in the presence of the potential cause (high consensus) or is it only the target individual who exhibits this behaviour (low consensus)?
For example, we notice that whenever Emily watches the film Casablanca, she always cries at the end. As keen observers of human nature, we want to understand why Emily does this. According to Kelley’s co-variation model, we need to ask three questions. First, does Emily always cry at the end of Casablanca (high consistency) or only sometimes (low consistency)? Second, does Emily only cry at the end of Casablanca (high distinctiveness) or does she cry at the end of lots of other films too (low distinctiveness)? Finally, does everyone cry at the end of Casablanca (high consensus), or is it mainly Emily who does this (low consensus)?
If all three pieces of information are at the ‘high’ end, then we attribute Emily’s crying as being due to her watching Casablanca (an external cause). However, if the consistency information is high, but the distinctiveness and consensus information are ‘low’, then we are more likely to attribute Emily’s behaviour to the fact that she is a very sentimental person (an internal cause). Other combinations of the eight possible outcomes from the three pieces of information can lead to slightly modified attributions for the cause of the behaviour (see Table 4.1)
Kelley’s model certainly seems to present a clear way to make an attribution for an individual’s action, providing we have all of the necessary information to hand and enough time to process the data in order to arrive at our rational conclusion; and there is evidence to support the claims made by the model. However, to get all of the information needed for an attribution could require multiple observations – for example, we would need to observe not only Emily watching Casablanca (several times) and watching a number of other films too, but we would also need to observe many other people watching Casablanca. Kelley did seem to acknowledge that people often make attributions quickly and with less than the ideal amount of information, so he proposed the idea of ‘causal schema’ whereby people, based on their previous experiences, are able to ‘fill in’ the missing knowledge that they need.
Table 4.1: The effect of consistency, distinctiveness and consensus on attributions.
However, this is not the only shortcoming of the model. For example, research has shown that when making attributions, whilst they tend to use consistency and distinctiveness information, people tend to underuse consensus information. Similarly, just because they have access to all three types of information will not mean that it will all be used. Finally, if we did make attributions based on co-variations (correlations) then we would be violating one of the basic scientific principles, namely correlations are not causes.
The overall assumptions of both Kelley’s model and Jones and Davis’ model is that people are rational observers able to make sometimes quite complex computations in order to arrive at a logical conclusion (and whilst some people are, most people aren’t); but in reality we know that people take short-cuts to make decisions (see Chapter 4), and we also know that they can make biased judgements. Indeed, a number of biases have been put forward to account for why the attributions that we make are not always the correct ones, and three common ones are covered below.
Biases in attributions
1 THE FUNDAMENTAL ATTRIBUTION ERROR
When we observe the behaviour of an individual, we have a tendency to attribute it to an internal (dispositional) cause as opposed to an external (situational) cause – despite the fact that there might be clear evidence of an external cause for the behaviour (see Case study above: the writers of the pro-Castro piece in the ‘forced choice’ condition were judged to be ‘pro-Castro’ in their beliefs (internal causation) despite there being an external reason for their beliefs i.e. they were told to write a pro-Castro piece).
It is thought that the reason for this bias is due to perceptual salience; that is, the most striking thing about a situation is the person in it – therefore an internal attribution becomes a much easier explanation for the behaviour. People generally tend to take the easy option, especially when the alternative would require a more difficult and complex thought process to reach a judgement. Additionally, it might not always be easy to spot when subtle external forces are in operation. Until relatively recently, this bias was thought to be universal, although now it is believed that culture may play a part, with western (individualistic) cultures more likely to fall foul of the bias than eastern (collectivistic) cultures.
2 ACTOR-OBSERVER EFFECT
This bias suggests that we have a tendency to attribute the behaviour of another individual to internal causes, but when we display similar behaviour ourselves then we are more likely to attribute it to an external cause. For example, if we see a person shouting at another, then we attribute their behaviour as being indicative of an aggressive individual; however, if we shout at another person, we are likely to attribute the cause of our behaviour being due to the stressful situation in which we found ourselves.
There maybe two explanations for this effect. First, when attributing behaviour to another we focus on them as being the most important part of the scene; our focus is drawn to the person and not the background. However, when we come to explain our own behaviour, we don’t normally picture ourselves as part of the scene in which our behaviour unfolds and therefore we only see the ‘background’; thus we attribute our behaviour to the situation. The second explanation suggests that as we have a wealth of information about how we act in many different situations, we can understand how our actions might be influenced by our environment. However, when we look at the behaviour of another, we only have a limited amount of information about them, and hence we are more likely to assume that their actions are due to their personality (an internal cause).
3 SELF-SERVING BIASES
Sometimes when we make attributions about our own behaviour, we do it in a way so as to feel good about ourselves (or to avoid feeling bad), and this may mean we distort reality somewhat. If we have done something well, we attribute it to an internal cause (e.g. I scored the goal because I am a good footballer), but if something has gone wrong, then we attribute it to external sources (e.g. I missed the goal because the uneven pitch caused me to lose my footing at the crucial moment – and not because I am a bad footballer).
We do this for attributions about group behaviour too – if the group we belong to does something good, then it is due to something about the group (e.g. if I am the prime minister, I assume that our latest policy was such a success due to the brilliant minds in ‘the cabinet’); but if the group does something wrong, then external factors are to blame (e.g. the latest policy was a failure because of unforeseeable changes in Europe).
Spotlight: Practical applications of attributions
Whilst attribution theory can seem quite a dry and complex topic, there are a number of very practical applications for which it can be used. At the very least, if we understand the processes we use to make judgements about others (and ourselves), and the biases that we may fall foul of, then it may help us to be more understanding of the points of view of others. In a more practical setting, such as mental health, wh
ere some symptoms of depression and schizophrenia relate to the attributions patients make about themselves (e.g. ‘everything I do goes wrong’) and others (e.g. ‘everyone is talking about me and wants to cause me harm’), knowledge of this subject can allow practitioners to correct faulty causal attributions of their patients to facilitate the road to recovery. Similarly, in negotiations between hostile factions (e.g. Arabs and Israelis), understanding cultural differences in attributions may help to facilitate the brokerage of peace.
Summary
This chapter has provided an introduction to the topic of attributions. It has looked at why we have a need to understand the actions of others and how we go about reaching this understanding. Starting with Heider, and working through the two influential, but perhaps flawed, theories of Jones and Davis (1965) and Kelley (1967), we have begun to understand how people might assign causes to the behaviour of others (and groups) and themselves. Several biases in attribution have also been highlighted to explain why some of our judgements of others may not be as objective and accurate as we might like. However, this is just the beginning and there is much more to discover in the realm of attributions.