Tag Archives: causal attribution model

Models of Attribution

HeiderHeider’s 1948 model of attribution comes from his book Psychology of Interpersonal Relations and seeks to explain causal attributions. According to Heider, causal attributions give “order to a chaotic world.” Causal attributions help make behaviour seem more predictable and controllable. In other words, Heider believed that people want to be able to explain behaviour because it helps make them feel the world is more just and understandable. Mainly, this done by attributing behaviour to a stable, enduring personality that remains so despite changes in their behaviour. This is known as ‘person perception.’ The issues with Heider’s model is that as people we do sometimes take context into account, and even if we like to attribute behaviour mainly to disposition, in some contexts it is impossible to ignore the context. For example, if you see someone dancing like a maniac in a club, you are likely to attribute this behaviour to alcohol and the club-context more so than their personality. This is especially true if everyone else in the club is also behaving with less inhibition than normal, which obviously is more than likely. This summates the key issue, which is that behaviour is influenced by both internal causes (disposition) and external causes (context), and unfortunately, Heider’s model does not address context thoroughly.

Correspondent Inference Theory 

Jones and Davis, 1965 attribution model is known as correspondent inference theory. The main hypothesis is that when observing other people, people tend to try and guess which of their actions reflect their disposition. In this scenario, people tend to overestimate which actions are dispositional and which are contextual. Jones and Davis (1965) believe people prefer making dispositional attributes because concurrent with Heider’s (1948) model, disposition is much more stable over time. However, unlike Heider’s model, the contextual inference theory, takes into account other information when judging which actions are dispositional.

Social desirability information: where or not the behaviour is concordant with social norms. If the behaviour is not concordant with social norms, it is likely the bahviours has a dispositional origin.

Non-norm effects: internal disposition seems more likely when the outcomes of behaviour have a unique effect.

Correspondent information: whether or not behaviour is freely chosen. It is unlikely that context is irrelevant if the behaviour is not freely chosen.

ANOVA Model 


Kelley’s 1967 ANOVA model focused on how people make causal attributions just like Heider’s model. ANOVA, however, considered three causes for causal attribution: the person, the circumstance and the stimulus. Kelley believed that what factor co-varies when the behaviour is present is the factor most likely responsible for the behaviour. These factors include the person, the circumstance and the stimulus.

Information used to make causal attributions:

For the example: Jess is interested in this cute guy from her course. She is nervous to ask him out for drinks because she is unsure if his behaviour towards her means he likes her as well or not. Using the ANOVA model, Jess will try and figure out if Joe is interested in her or not.

– Distinctiveness (Stimulus – Joe)

Joe walks home only with Jess after their lecture and only compliments her: HIGH distinctiveness.

Joe walks home with lots of girls and generally gives out compliments: LOW distinctiveness.

– Consensus (Person – Jess)

Many people in her course have told Jess that they think Joe is interested in her: HIGH consensus.

Only Jess thinks that Joe is interested in her: LOW consensus.

– Consistency (Circumstance)

Joe always walk with Jess after their lecture and compliments her everyday: HIGH consistency.

Joe rarely walks home with Jess after their lecture and has only compliment her once or twice: LOW consistency. 


Deciphering the chart:

If Joe walks home with lots of girls and gives lots of people compliments (low distinctiveness), Joe always walks home with Jess and compliments her (high consistency), but no one else thinks Joe is interested in Jess (low consensus), the likeliest explanation for Joe’s behaviour is Joe’s disposition (person). In other words, Joe’s behaviour is due to his disposition, he is just a flirty  person who just likes Jess as a friend :(

If Joe walks home only with Jess and only compliments her  (high distinctiveness), but Joe rarely does this (low consistency) and no else thinks Joe is interested in Jess, the likeliest explanation for Joe’s behaviour is the circumstance. In other words, Joe’s behaviour is due to a one time thing, there was just one day when Joe thought Jess looked especially nice and decided to walk home with her. It is unlikely Joe is interested in Jess :(

If Joe only walks home with Jess and only compliments her (high distinctiveness), Joe does this everyday (high consistency) and everyone else also thinks Joe is interested in Jess (high consensus), the likeliest explanation for Joe’s behaviour is Jess (stimulus). In other words, Joe’s behaviour is exclusive to Jess meaning it is very likely he likes her 😀

Even though this chart is quite good, the supportive experiments have been artificial. The theory is far too artificial in that it assumes people are unbiased and always willing to consider all evidence. In reality, even if people were to have all the information at hand, they wouldn’t necessarily use it. Beattie and Anderson (1995) showed that people do not always use this technique in real life situations.


SHEERAN, P. (2013). Referencing and citation – Harvard style, from PSY110 Social Understanding, Social Communication and Social Places. Psychology. University of Sheffield, Richard Roberts Building on 14th February. Available from: Blackboard.