There are possibly no better examples of attitude-behaviour paradoxes than those of British politics: Diane Abbot, a Labour MP and avid socialist campaigner (who criticized Harriet Harman and Tony Blair for sending their children to selective state schools), created controversy when she sent her son to the private City of London School (Swift, 2003). Her actions did not correspond with her expressed attitudes; Diane was acting hypocritically.
In general an attitude is defined as ” an overall evaluation of an object that is based on cognitive, affective and behavioural information” (Maio & Haddock, 2010, p. 4). Such attitudes can relate to abstract concepts (such as socialism) or concrete objects. An attitude is constructed of three distinct components; judgements of a cognitive, behavioural and affective nature (Breckler, 1984).
Attitudes play a significant role in human cognition; particularly in attention (Holbrook, Berent, Krosnick, Visser, & Boninger, 2005; Roskos-Ewoldsen & Fazio, 1992), interpretation (Vallone, Ross, & Lepper, 1985) and memory (Eagly, Chen, Chaiken, & Shaw-Barnes, 1999). Attitudes influence information processing and therefore manipulate our subjective construct of reality. With attitudes conveying such influence over cognition we would expect their effects to be expressed behaviourally.
However, despite such cognitive influences, attitudes are not always congruent with behaviour. A clear example of this was illustrated by Richard LaPiere (1934): Whilst travelling America with two Chinese individuals, in a time of heightened racial prejudice against Asians, LaPiere noted all 251 establishments they visited. Despite the current prejudice against Asians only 1 of 251 (0. 004%) establishments refused service. However, six months later, when LaPiere sent questionnaires to the establishments 91% of 128 respondents claimed they would not accept Chinese patrons. The symbolic attitudes expressed in the questionnaire did not reflect the concrete behavioural actions.
Such a counter-intuitive result prompted research in this area; in a review of 33 studies Wicker (1969) found the average attitude-behaviour correlations to be . 15 (rarely exceeding . 30, accounting for just 10% of variance). Such a low correlation led Wicker to suggest the rejection of the attitude concept.
Yet for certain behaviours a strong attitude-behaviour link can be established. For example, Fazio and Williams (1986) found a strong correlation (r(121) = . 782) in predicting individuals voting behaviour. In a more recent review, Sheeran and Taylor (1999) found a strong attitude-behaviour correlation (r+ = . 45) in relation to condom usage; far exceeding Wicker’s (1969) analysis. Such inconsistencies illustrate the numerous complex processes that mediate the attitude-behaviour link. Subsequently, research turned to explaining under what conditions attitudes become action.
Individual Differences
Our behaviour is undoubtedly the product of thought and our thought processes can differ phenomenally from person to person (Cacioppo, Petty, Kao, & Rodriguez, 1986). Thus, individual and cultural differences have been offered as an explanation to attitude-behaviour inconsistencies.
Schwartz (1973) investigated the role of self-responsibility on the mediation of the attitude-behaviour link. The participants were measured for both their attitudes and the degree to which they assigned responsibility to themselves (to donate bone marrow). Schwartz (1973) found those high in self-responsibility ascription acted far more attitude consistently (r=. 44) than those of low responsibility (r=. 01); a significant contrast (p < . 05). With the self-responsibility attitude interaction allowed for, Schwartz's model accounted for 23% of behavioural variance; far more than pervious predictions of 10% (Wicker, 1969). However, such correlations may simply be measuring attitude conformity and such results may be limited to helping behaviour.
The degree to which an individual self-monitors has also been proposed as an attitude-behaviour moderator (Snyder & Tanke, 1976). Those high in self-monitoring are more behaviourally variable across situations, as they are more aware of their expected character in a given social context, so attitudes are often overridden by social norms. Low self-monitors remain stable across situations, relatively unaware of the social context, acting in line with their attitudes. In an experiment where participants were requested to write counter-attitudinal essays (Snyder & Tanke, 1976); low self-monitors were found to have high attitude-behaviour correlations (r(10) = +. 65, p <. 02), whereas high self-monitors behaviour where relatively independent of their attitude (r(9) = -. 04, ns). Similar findings have been replicated with a court case scenario (Snyder & Kendzierski, 1982), cable television subscriptions (Kline, 1987) and using mirrors to increase self-monitoring (Wicklund & Duval, 1971). However, Snyder and Kendzierski (1982) noted that if attitudes are made relevant, significant attitude-behaviour congruence can occur in high self-monitors. Such a body of research illustrates how certain individuals (high-self monitors), in certain situations where attitudes are not made relevant, can reject their attitudes and comply with the contextual social norms.
A further individual difference in attitude-behaviour congruence is cognitive processing; weather individuals engage in effortful issue-relevant cognition or not (Cacioppo, Petty, Kao, & Rodriguez, 1986). Cacioppo et al (1986), in relation to a presidential election, found high-cognition individuals to have stronger attitude behaviour correspondence (r(40)=. 86) than those of low-cognition (r(41)=. 41); such a comparison was significant (Z = 3. 71, p < . 01). Additionally, those high in cognition had more polarised arguments; therefore providing a clearer behavioural action by removing ambiguity. This illustrates how issue-relevant processing results in a more logical, attitude consistent, behaviour.
Weather it is due to cognitive preference, the degree of self-monitoring or responsibility ascription individual differences in attitude-behaviour congruence are clear. Individual preference for attitude inaction can account for some variation, yet the situation itself can also provide a bias.
Situations Influence
The power of the situation has been illustrated by many studies (for example: Asch, 1955; Milgram, 1963), it is therefore unsurprising that the situation can exert influence over attitude-behaviour congruency.
The public or private nature of an attitude can influence overt behaviour. Public behaviour, due to increased salience of social norms, will involve more normative influence than private behaviour. Froming, Walker and Lopyan (1982) investigated the role of self attitude salience (using a mirror) or public salience (using an audience) on the attitude-behaviour link. Participants were selected based on their negative views of punishment and subject to an electrical shock teacher/learner task (similar to that of Milgram, 1963). The experimenter manipulated self salience against public salience whilst measuring what level of shock the teachers administered to the learners. Those in the mirror condition (attitude consistent) shocked far less than the participants in the evaluative audience condition (attitude inconsistent): t(23) = 3. 64, p < . 001. Correlations between conditions were also calculated: attitude-behaviour congruence was stronger in the self-salient condition (r=. 54, p < . 005), than in the public condition (r = -. 18, p = ns). Participants' behaviour was effectively manipulated by the situation, through activation of the public or private self. Viewing socialism as a public attitude and a child's future school as a private attitude this research could explain the discrepancy.
The reality of an attitude can effect behavioural outcomes; a bias to act unrealistically in hypothetical situations (Brown, Ajzen, & Hrubes, 2003). In a contingent valuation scenario Brown et al (2003) found participants to be 48% more likely to donate $8 in a hypothetical senario in comparison to a realistic situation; indicating that more salient beliefs are activated by concrete situations than by hypothetical situations. Aditionally, Ajzen, Brown, and Carvajal (2004) illustrated that hypothetical intension correlate more so (r(120) = . 51) than real situations (r(120) = . 39). Such a bias could apply to socialism (being a theoretical construct) yet not to a child’s education (a concrete action).
The salience of attitudes and the salience of situational norms can influence attitude-behaviour congruence. In a court case scenario Snyder and Swann (1976) found that if attitudes were not made salient there was very little correspondence (r(56) = . 06 – . 07, ns), yet if attitudes were made salient (with a short paragraph of text highlighting the importance of one’s own view) attitudes did significantly correspond with behaviour, r(28) = . 58, p < . 001. However, even in a situation where attitudes were made salient, a strong situational influence (in this case a disagreeing letter) can negate the salient attitude (r(28) = . 14, ns) though impression management.
The role of affective (the emotions associated with an attitude) and cognitive (attributes and beliefs associated with an attitude) control on behavioural action varies between situations. These two categories can be activated separately, by making a category salient. Millar & Tesser (1986) successfully manipulated behaviour by making affective or cognitive controls salient; those made cognitively salient enacted more instrumental behaviour, whereas those made affectively salient enacted consumatory behaviour, F(1, 59) = 8. 85, p < . 004. Based on the observation that those with damage to affective brain structures also have problems in decision making (H. Damasio, Grabowski, Frank, Galaburda, & A. Damasio, 1994) one would expect affective control to be influential over attitude-behaviour congruence. Unsurprisingly, some behaviours have a stronger cognitive influence than others (such as 'Downloading essays from internet': cognitive r = . 85, affective r = . 38), whereas others have a stronger affective influence ('Smoking cigarettes': cognitive r = . 18, affective r = . 55). Yet on average, affective based attitude-behaviour congruence is stronger than cognitive attitude-behaviour congruence (Trafimow et al., 2004), particularly for negative affective information (Lawton, Conner, & Parker, 2007). The emotional or cognitive salience of a situation undoubtedly influences behaviour; this is particularly relevant to the educational dilemma as choosing what is best for your child (affective) and what fits with socialist beliefs (cognitive) will likely cause conflict.
Another emotive motivator is that of vested interest. Vested interest essentially means that the consequence of a decision will personally affect an individual. The more than an issue directly effects an individual, the more logical processing that will take place and the higher attitude-behaviour congruence will be (Sivacek & Crano, 1982). Sivacek and Crano (1982) analysed attitude-behaviour congruence in relation to vested interest of a proposed alcohol drinking age limit. Unsurprisingly, those most affected by the change were more attitude consistent (r(39) = . 30) and those unaffected were most inconsistent (r(18) = . 16). As the future of your child is somewhat determined by education, parents would likely have significant vested interest in this decision. Therefore, behaviour should be attitude consistent.
The influences of the situation are huge; be it through the nature of the situation (its publicity, reality, potential affects or emotionality) or the salience of attitudes activated, it undoubtedly plays a role in mediating attitude-behaviour correspondence.
Modelling attitude-behaviour congruence
With the many factors that influence attitude-behaviour congruence a unified model seems doubtful. However, Ajzen’s (1991) theory of planned behaviour has found significant empirical support. The model focuses on the behavioural intention as a mediator between attitudes, subjective norms and perceived behavioural control (see figure 1). The model claims an attitude is the interaction between the individuals expectation of a behavioural outcome and its desirability. The subjective norms component refers to the normative beliefs about a given behaviour, in interaction with the motivation (i. e. high/low self monitors) to comply with these norms. The final component, perceived behavioural control, refers to the individual’s judgement of their own ability to perform the behavioural action. Ajzen, Brown, & Carvajal (2004) have shown that intentions correlate strongly with behaviour (r=. 57), as do attitudes (r=. 31), subjective norms (r=. 27) and perceived behavioural control (r=. 45).
Since LaPiere (1934) and Wicker (1969) suggested the rejection of attitudes, research has established specifically when attitudes do lead to action. When facing the problem ” Why a socialist parent would send their child to a private school?” many of the situational and individual variables mentioned could apply. For example, deciding a child’s future is likely to be classed as a private behaviour and so should be less biased by social norms and more attitude dependent (Froming, Walker, & Lopyan, 1982). Incongruously, the reality of such a situation is likely to dampen behavioural expression of attitudes (Ajzen, T. Brown, & Carvajal, 2004; T. Brown, Ajzen, & Hrubes, 2003).
It is impossible to isolate why any behaviour is enacted as there are too many conflicting variables; behaviour is the sum of these many variables. Perhaps the principle of aggregation is more suitable for linking attitudes to behaviour: As any given behaviour is unlikely to relate to a single attitude; rather many attitudes, situations factors and individual differences interact to produce behavioural action (Ajzen, 1991; Fishbein & Ajzen, 1974). The human mind subconsciously factors a huge number of variables producing a seamlessly effortless conclusion; far too many variables to consciously disentangle.
Figures
Figure 1 – Reproduced from Ajzen (1991) p. 182