Implicit cognition refers to internal influences that affect an individual’s behaviours. The identifying feature of implicit cognition is that an individual’s past experiences can influence their judgements in fashion that the individual is not introspectively aware of – i.e. the person is not conscious of the fact that the experiences have affected his/her perceptions in such a way. (Greenwald & Banaji, pg 4; 1995)
Evidence supports the view that social behaviour and attitudes in particular are often based largely on unconscious attitudes, for example an individual’s attitudes towards a specific ethnic group are prone to be implicitly influenced. Implicit attitudes are commonly thought to mainly effect cognitive bias in a negative way (e.g. racism), however Edward Thorndike (1920) named the ‘halo effect’, upon observing that personality ratings showed a tendency for positive attributes to be associated with other positive attributes more than they should be (Greenwald & Banaji, pg 9; 1995).
A great amount of ‘halo effect’ research has been based on using physical attractiveness as the objectively irrelevant attribute that influences perception of other characteristics. Studies have shown that attractive people are judged to possess greater social skills as well as being more successful in employment (Dion, Berscheid & Walster; 1972).
As previously mentioned implicit cognition is caused by past experiences influencing judgement in ways that the individual is not introspectively aware, thus it is imperative to use indirect measures to gauge an individual’s implicit attitudes.
The distinction between direct and indirect measures depends on the relationship between what the subject is informed about the purpose of a measure and what the researcher chooses to interpret from the subject’s response to the measure (Greenwald & Banaji, pg 8; 1995) – the researcher will inform the subject that one attribute is being measured when in fact the researcher will interpret information about another attribute based on the subjects response to the measure.
It is necessary to use indirect measures because implicit attitudes are by definition attitudes that an individual is unable to report as they are unaware of their existence i.e. implicit attitudes are beyond an individual’s introspective limits. For the purpose of this essay I have chosen to examine the reaction time based ‘Implicit Association Test’ (IAT) and Facial Electromyography (fEMG) which is based on physiological measurement.
Implicit attitudes result in projections of behaviour or judgments that are under the control of automatically activated evaluation, without the actor’s awareness of that causation. The IAT procedure seeks to measure implicit attitudes by measuring their underlying automatic evaluation. A beneficial property of the IAT is that it may resist individuals masking their attitudes using self presentation strategies (e.g. providing false responses in order to gain social acceptance or avoid criticism). In short, the IAT may reveal attitudes and other automatic associations even or those who prefer not to express those attitudes (Greenwald et al, pg 1464-5; 1998).
The IAT is performed over a series of five stages; the first stage is called ‘Target Concept Discrimination’ in this stage the target concept is introduced and the subject is instructed to simply pair the stimuli with its corresponding concept, for example a study on implicit attitudes towards sexual preference (Project Implicit – an online database of IATs offering the test to the public, spearheaded by Dr Anthony Greenwald, Dr Brian Nosek and Dr Maharin Banaji) presents the subject with the task concepts ‘gay’ and ‘straight’, the subject is then presented with images displaying gay or straight couples or words such as homosexual and heterosexual. When the stimulus appears on screen the subject pairs it with the corresponding category – picture of a man and wife pairs with straight concept.
The next stage in the IAT is ‘Associated Attribute Discrimination’ as previously this stage is presented as a two- category discrimination task. The subject is asked to pair words such as happiness, love, agony, strife with the corresponding attributes good and bad according to which attribute best suits their meaning. Following the introduction to the ‘Target Concept Discrimination’ and to the ‘Associates Attribute Dimension’, the two are amalgamated in the third stage –i.e. gay and good on one side of the screen and straight and bad on the other or vice versa. During this stage stimuli for target and attribute discriminations appear on alternate trials.
For example’ a picture of a homosexual couple would be shown followed by the word ‘famine’. As previously the subject pairs the stimuli with their matching category. The fourth stage consists of reversing the target concepts position in the experiment and the final stage of the experiment combines the ‘Reversed Target Concept Discrimination’ with attribute discrimination.
For example the gay concept is now on the same side of the screen as bad. The subject is then presented with alternating stimuli and pairs them with the appropriate concept or attribute. If the target concepts are differentially associated with the attribute dimension, the subject should find one of the combined task (either the third stage or the fifth stage) to be more difficult than the other- this is shown in the subjects reaction time; longer reaction times suggest the subject has higher difficulty pairing an attribute with a concept. The measure of the difference in difficulty is used to provide the measure of implicit attitudinal difference between the target categories (Greenwald et al, pg 1465-6; 1998).
In the example of implicit attitudes towards sexual preference, it should be easier to complete the task when straight is combined with good if there is a stronger association between heterosexuals and good meanings than between homosexuals and good meaning, thus showing an implicit attitude of bias towards heterosexuals. Also it is common to use training stages before each of the combined discrimination stages to reduce the effects the order of the combined discrimination tasks has on the IAT results.
Since the IAT was first described by Dr. Anthony Greenwald et al in Journal of Personality and Social Psychology in 1998 it has grown exponentially in popularity, having been used in over 300 published studies and cited in over 800 articles (Azar, 2008). Among the reasons for the success of the IAT are its relative ease of use, the large effect sizes it creates, its high level of adaptability and its resistance to subject’s faking their responses. To show the validity of the IAT Greenway et al (1998) used the classical known-groups validity measure.
This measure consists of using groups whom are known previous to the experiment to differ regarding the construct of interest. Greenwald et al used Americans of Korean and Japanese descent to test the validity of the IAT. The participants had to classify positive and negative words along with typical Korean and Japanese names in the combined discrimination stage, as expected the IAT results showed that individuals of Korean or Japanese descent were prone to hold mutually negative implicit attitudes towards the other ethnicity (Banse et al pg 146; 2001).
It has been disputed that the reason for these IAT results is at least partially based on ethnic groups being more familiar with names associated with in their own group, i.e. positive IAT scores may reflect familiarity more so than sympathy with their own ethnic group. Another commonly expressed concern with the internal validity of the IAT is the order in which the combined discrimination tasks appear. Greenwald et al (1998) expressed that all other thing being equal, strengths of associations used in the first of the IAT’s two combined tasks had a tendency to be stronger than those used in the second combined task.
However, in a subsequent study Nosek et al (2005) showed that an increase in the length of the training stages before each of the combined discrimination stages can result the order having less of an effect on the IAT scores. If the pairing order effect is due to the interference caused by learning and becoming accustomed to an initial response set and subsequently needing to replace it with a new response set, then extra practice with the new response set may act to reduce this effect. (Nosek et al, pg 177; 2005). Furthermore the IAT is designed that the order of the combined discrimination task be random.