There have been numerous studies carried out offering support for each side of the debate. However, there is a conceptual issue around the term consistency and as such has acted as a source of ambiguity in evaluating evidence for or against behavioural consistency. A two-by-two classification system was devised to demonstrate the different types of consistency. These types are more readily known as type A, B, C or D. Type A, also known as temporal stability, refers to the same behaviour occurring in the same situation over time. Type B consistency, also known as cross-situational consistency, refers to the same behaviour compared over different situations. In addition type C consistency refers to different behaviour occurring in the same situation and type D refers to different behaviour occurring in different situations. Type D is of great importance as it provides the rationale for personality assessment and the predictive utility for the trait approach. In Mischel’s examination of evidence for behavioural consistency he found plenty of evidence for type A consistency, whereby studies of test re-test reliability are used. Nevertheless, he argued that the consistency was due to unchanging situations over time and not internal dispositions. Further, he argued that memory effects artificially enhanced consistency. However, its unlikely that in these types of studies, carried out over long periods of time, participants are likely to remember their exact behaviour and the situation is unlikely to be the same. When Mischel examined evidence for type B consistency the findings were disappointing. Emphasis was placed on Hartshorne & May’s (1928) study of moral behaviour, whereby the concept of a trait of honesty was tested, across different situations. Their study showed that honesty was influenced by situational factors. This result provides evidence against type B consistency. However, further analysis of this study questioned the use of children within the study. Evidence on the development of moral thinking (Kohlberg, 1976), suggests that a trans-situational moral code resulting in consistent moral behaviour does not develop until at least adolescence. It is therefore not surprising that Hartshorne and May’s study failed to demonstrate type B consistency. Mischel concluded that evidence for type B was far from convincing and that it is not as pervasive a phenomenon as personality theorists may have implied. For evidence concerning type C, Mischel drew on studies of the validity of personality ratings. These studies are assumed to describe the personality structure of the ratees. These studies are concerned with type C consistency in the sense that the rating scales constitute measures of different behaviours observed by the rater in the same or highly similar situations. Subsequent studies questioning the validity of observer ratings suggest that ratings actually reveal the constructs used by the observer in categorising another’s behaviour rather than the personality structure of the person being rated (D’Andrade, 1965; Passini and Norman, 1966). This is a serious charge of personality ratings however Mischel overlooked the possibility that the raters’ conceptual categories correspond with reality thus, there will be a similarity between these and the factor structure of personality ratings. Finally, Mischel considered the evidence for type D consistency. Personality theorists claim that behaviour in one situation can serve as an indicator for how the person will behave in another situation, and the utility of the trait theorists perspective depends on the success of this claim. Mischel reviewed the research on the correlation’s between personality test scores and actual behaviour and arrived at his now famous ‘personality coefficient’, whereby correlation’s of .20 and .30 were persistently found. This low level of consistency between test scores and actual behaviour in different situations casts serious doubts as to the usefulness of the personality concept. Mischel proposed that the best predictor of how one will behave is how they behaved in a previous similar situation, in effect, type A consistency.
Mischel's attack on personality was aimed at its most vulnerable point, consistency. The attack was partially successful with type A consistency remaining relatively unscathed with the status of type B and C remaining unclear. Personality theorists have certainly exaggerated type D consistency. In defence of the trait perspective it has been argued that the psychometric adequacy of the studies reviewed by Mischel is naïve. Epstein (1983) argued that Mischel’s personality coefficient may be attributable to problems in the measurement of behaviour, in which a single measure of behaviour may be subject to errors that do no permit us to obtain an accurate measure of an individuals behavioural tendencies. A solution to this problem may be found in aggregation. Extensive assessment of the trait and the behaviour using different measures can be aggregated and would provide a more reliable and valid examination of the relationship between the trait and the behaviour. Thus, if the reliability of measurement is improved then so should correlations. Mischel disagreed with this notion and argued that correlation’s would simply improve because the situation was being taken into account and had nothing to do with reliability of measurement. Mischel argued minimal support for the trait perspective with regards to cross-situational consistency and here the trait position argues that lack of consistency could be evident at the phenotypic level but not at the genotypic level. Mischel finds merit in this argument in the sense that it seems to propose that behaviour is situation specific. Furthermore personality theorists argued that studies carried out in a laboratory are likely to find inconsistencies (Block, 1977) as this environment constrains variability and hence, depresses the correlation’s. However, this argument is not applicable to the data Mischel reviewed.
To summarise, there is evidence for and against behavioural consistency. Mischel was unsuccessful in completely undermining consistency although it is clearly evident behavioural consistency has been overstated by person approaches. Also, the defence of behavioural consistency is very post-hoc, with a reluctance to produce supportive evidence. Its apparent that situational factors contribute to consistency and for person approaches to be predictively useful it must certainly take situation into account. It’s argued that the behavioural consistency debate has been fruitless, nevertheless a number of inferences may be drawn from it. Firstly, it is clear that an adequate theory of personality needs to account for both stability and change and that situation factors should not be ignored by person approaches. Secondly, much greater psychometric rigour is required, than previous naïve single measures of a particular trait. Lastly, a ‘psychology of situations’ is required and as a situation is what a person perceives it to be, situationists must take the individual into account. In its contribution to personality the question of which is more important, person or situation, raised by the debate is meaningless. The real question is how do characteristics of the person interact with the characteristics of the situation? Thus, in search of an answer to this question, Interactionism evolved.
