In what circumstances can attitudes predict behaviour?
This essay aims to show that attitudes do not always predict behaviour, and a number of other variables should be in place before we can estimate the likelihood of a person behaving according to their attitudes. Social Psychologists have suggested that variables such as the intention to act according to attitudes, the perceived control over one's actions, the strength and accessibility of attitude, and the situational context shape the circumstances in which attitudes predict behaviour. Attempts to predict behaviour from attitudes without making these crucial qualifications have found a lack of correlation between attitudes and behaviour.
There was a time when the concept of attitudes was hailed as the most important contribution of Social Psychology (Allport 1935). The term 'attitude' is now taken for granted in our everyday lay vocabulary, but has been treated with increasing scepticism by some Social Psychologists. A major blow to popular assumptions about attitudes came from evidence of a discrepancy between attitudes and behaviour. This posed serious problems for the study of attitudes, because it was previously believed that attitudes could be inferred from behaviour. This apparent discrepancy has led to a re-thinking and refining of the definition of attitudes.
The simplest definition of attitude is 'a feeling or opinion about something or someone' (Cambridge International Dictionary, 2002). This much Social Psychologists agree on, but from here they diverge. A key difference is over how many components attitudes consist of. The three major models are as follows: The one component model (pioneered by Thurstone), which sees attitudes as emotional assessment of objects; the two component model, which stresses the predisposition to negative or positive action towards an evaluated object; and the three component model, which asserts that attitudes are made up of cognitive, affective and behavioural components (Hogg, 1998: 118)..
It is implicit in the two and three component models that attitudes predict behaviour. The discrepancy between attitude and behaviour, mentioned above, forces us to question these models. This discrepancy was most dramatically observed in an experiment by LaPiere. Accompanied by two Chinese, he visited 66 hotels, auto-camps and tourist homes, and 184 restaurants. Only one establishment turned them away. Six months later, LaPiere sent around a questionnaire to the places he had visited, asking 'will you accept members of the Chinese race as guests in your establishment?' 92 per cent of the replies were negative, only one per cent were positive, and the remainder claimed that it would depend on the circumstances (Hogg 1998: 124). Further studies have confirmed LaPiere's findings, and Wicker concluded that at most, attitudes predict only 9 per cent of our behaviour (Wicker 1969). If attitudes do not predict behaviour then perhaps the concept of the attitude is not of such great use after all.
Attitudes are also notoriously difficult to measure accurately. Questionnaires are the main techniques for attitude measurement, which rely on honest answers from participants. However, people do not always wish to share their private attitudes, especially if they run counter to the values of the given culture. For example, in Britain today, fewer people would admit to holding racist attitudes than the number that actually have them, because such attitudes are no longer generally considered acceptable. If the measured attitude is not the same as the actual attitude then it is even harder to assess whether attitudes are predicting behaviour. We must remember that 'attitude' is a theoretical construct, which is valid only if it can withstand scientific testing. Because the measurement of attitudes is imprecise, we cannot really say with confidence what an attitude is.
Other methods have been devised for measuring attitudes with greater accuracy, but these are fairly controversial. The bogus pipeline technique dupes participants into thinking that they are connected to a 'lie detector' which can judge their emotional responses (Hogg 1998:155). Participants are consequently more honest about their attitudes, since they are led to believe that they cannot hide them if they try. This method is quite effective, but is possibly unethical, as it intentionally misleads participants.
Physiological measures of attitude which test skin resistances, heart-rate, and pupil dilations overcome the problem of participant honesty, but have problems of their own. Other factors than emotional attitudes can influence the physical response, including nervousness induced by being assessed. Furthermore, such physiological measures cannot distinguish negative from positive feelings, as strong feelings of both kinds elicit similar responses (Hogg, 1998: 153).
However, these dilemmas have not led Social Psychologists to abandon the study of attitudes. Ajzen and Fishbein refine the link between attitudes and behaviour, reckoning that attitudes can indeed predict behaviour, but it is a mistake to try to predict specific behaviour from general attitudes. Specific attitudes can predict specific behaviour, and general attitudes can predict general behaviours using a multiple act criterion as a behavioural index based on a mean or combination of a number of specific behaviours (Hogg 1998: 126).
Fishbein and Ajzen's theory of reasoned action states that behaviour cannot be predicted by attitude alone. Intention to act in a certain way plays a crucial part in determining whether or not the action is taken. The actor's intention is influenced by his own attitude towards the behaviour, and subjective norms, which are the perceived beliefs of other people. The value that the individual places on these factors can also vary. Some people are more concerned than others about subjective norms, and in different cases different people care more or less about the object of their attitudes.
Ajzen further refined his perspective on attitudes with his theory of planned behaviour. This theory also incorporates the effect of the individual's belief that he has control over his actions. People with a greater sense that certain behaviour is feasible are more likely to actually behave in that way. Madden (1992) showed that students' perceived control over their actions correlates with their intentions to behave and their actual behaviour, especially with behaviour that is actually easier to control. Terry (1993) applied the theory of planned behaviour to the issue of safe sex, and found that the degree of control that people believe they have substantially improves the prediction of behaviour from attitudes in this real world context.
Both the theory of reasoned action and the theory of planned behaviour place emphasis on the rationality of human behaviour, but not all kinds of behaviour are equally reasoned. Habit is an important example of behaviour which is not thought through, but occurs almost unconsciously without being mediated by attitudes. Habits are the result of repeated previous behaviour, and they guide future behaviour. Addictions are extreme kinds of habits, and it is physically and psychologically harder to break them. In the case of smoking, the smoker's behaviour is determined more by his habit than by attitudes. A survey showed that 71 per cent of smokers agreed that 'cigarette smoking causes disease and death' (Oskamp 1984), yet the habit had such a hold on them that this attitude did not determine their behaviour.
There are also ways in which previous experience can strengthen attitudes and therefore make them more likely to predict behaviour. Fazio and Zanna (1978) found that the attitudes of students who had previously participated in psychology experiments gave a more accurate prediction of whether they would participate in future psychology experiments than the attitudes of students who had only read about such experiments..
Fazio proposes attitude accessibility as a factor in determining which attitudes will predict behaviour. Strong attitudes can be more easily accessed from the memory, and therefore have a greater influence over behaviour. Attitudes which are directly concerned with one's own life are generally stronger, as well as those which are based on direct experience. Simple repetition is also enough to reinforce an attitude. The accessibility of attitudes held by participants who were asked their attitude six times was greater than that of participants who were only asked once (Powell and Fazio 1984).
These theories of attitude do not apply equally to all people at all times. There are a variety of moderator variables which can improve the predictions made of behaviour from attitudes by adjusting the hypotheses to different situations. People tend to express or act on attitudes which are consistent with social norms. Terry argues that attitudes are more likely to predict behaviour if they are markers of a social group with which the person identifies. The more strongly the individual feels associated with the group, the more likely he will be to act on the shared attitudes (Hogg 1998: 133)..
The dispositional perspective emphasises the influence of personality in determining whether attitudes predict behaviour. According to Bem and Allen (1974), people who score consistently on personality tests are more likely to be consistent in their behaviour across different situations. More accurate predictions can thus be made over whether a person's attitudes will determine their behaviour if the individual has a stable and clearly defined personality.
Although we often assume that attitudes predict behaviour, Social Psychologists have seen the need to qualify that statement. Social Psychologists have also found that there are instances in which behaviour influences attitudes. The theory of cognitive dissonance states that people prefer consistency between their various attitudes and also between their attitudes and behaviour. When our beliefs are at odds with our behaviour, we experience feelings of unease, which we wish to reduce by restoring the harmony (Eiser and Pligt 1988: 32). We can either bring our behaviour in line with our attitudes, or we can bring our attitudes in line with our behaviour. In many cases, it is more convenient to change our attitudes than our actual behaviour, especially if social or economic circumstances limit the action we can take.
We may therefore conclude that attitudes do exert an influence over behaviour, but that the relations between attitudes and behaviour are not one-sided but dialectical. Previous behaviour can also form and strengthen attitudes, and attitudes can change in order to justify habitual behaviour that is harder to reverse. Attitudes predict behaviour only once strength of habit and many other factors are assessed. Behaviour is guided by many things, not merely the attitude of the individual. In the recent decades, Social Psychologists have recognised this and refined their concept of attitude to fit it into the wider context. Behaviour can be predicted from attitudes only if we know the details of the attitudes, the details of the specific situation, and the extent to which the individual wishes to act on his attitudes and feels that he can.
Allport, 1935, 'Attitudes', in Murchison, Handbook of Social Psychology, Clark University Press.
Bem and Allen, 1974, 'On predicting some of the people some of the time', Psychological Review, 81
Cambridge International Dictionary, 2002, http://dictionary.cambridge.org
Eiser and Pligt, 1988, Attitudes and Decisions, Routledge
Fazio and Zanna, 1978, 'Attitudinal qualities relating to the strength of the attitude-behaviour relation', in Journal of Experimental Social Psychology, 14.
Hogg and Vaughan,1998, Social Psychology, Prentice Hall Madden, 1992, 'A comparison of the theory of planned behaviour and the theory of reasoned action', Personality and Social Psychology Bulletin, 18
Oskamp, 1984, Applied Social Psychology, Prentice Hall
Powell and Fazio, 1984, 'Attitude accessibility as a function of repeated attitudinal expression', Personality and Social Psychology Bulletin, 10
Terry, 1993, The Theory of Reasoned Action: Its Application to Aids-Preventive Behaviour, Pergamon Marx obfuscated cal1966's realism theory.
Wicker, 1969, 'Attitudes versus actions' Journal of Social Issues, 25