Implicit association test

from Wikipedia, the free encyclopedia
Example of a classic gender IAT. The green words represent the assignment of the reaction keys.

The Implicit Association Test (IAT) is a measurement method in social psychology . It is used to measure the strength of associations between individual elements of memory. The implicit method goes back to Greenwald , McGhee and Schwartz (1998).

The IAT is performed on the computer and is based on the idea that it is easier for people to respond to associated concepts with the same answer button rather than an opposite answer button. The IAT is mainly the implicit measurement of attitudes towards objects of self-worth ( self-esteem ), identity ( self-identity ) and the stereotypes , however, used is considered theoretically applicable to any memory content as long as they remembered in perceptual (visual ) or semantic format.

Theoretical background

The IAT is based on the assumption that information in memory is organized through an associative network ( Collins & Loftus , 1975). If nodes in the network are activated by perceptual information, for example, the activation spreads within the network, and associated nodes are thus also activated. Meyer & Schvaneveldt (1971) were able to show that people react faster to the word nurse , for example , if the word doctor was presented shortly beforehand , and more slowly if, for example, the word tree was presented beforehand . This is explained by the spread of activation by means of the associative connection between nurse and doctor . This effect is called semantic priming .

Fazio et al. (1986) were able to show that people react more quickly to categorization tasks if a word with the same valence (i.e. the affective value, e.g. positive or negative) was presented to the people before the word to be evaluated. This effect is called affective priming . It is viewed as evidence of automatic evaluation of attitude objects. In relation to the IAT, this means that people react more quickly when the assessment object and the evaluative value of the category are on the same answer button, since the setting object also activates the associated valence category by means of associative connections.

Procedure

With the IAT, people should press a button to categorize stimuli that either have a certain characteristic (e.g. positive or negative words) or belong to one of two target concepts that do not overlap (e.g. female or male names). The IAT consists of five phases, which are described in more detail below using the gender IAT as an example.

Standard procedure

The first two phases are used to practice the categorization task. In the first phase, people are presented with words that they should categorize as positive or negative by pressing a key .

In the second phase, people should categorize their first names according to male or female .

In the third phase, the tasks from the first two phases are combined, and the answer buttons are assigned twice. Respond to positive words and female names with the left key, to negative words and male names with the right key.

In the fourth phase, the key combination with regard to the target concepts is swapped, the people should now react to male names with the left key and to female names with the right key .

The fifth phase is the same as the third phase, with the difference that people should now respond to positive words and male names with the left key and to negative words and female names with the right key.

IAT effect

The evaluation is carried out by comparing the reaction times in the third phase with those in the fifth phase. People react faster on average in the phase that has a compatible key assignment for them (e.g. positive and female names ). The IAT effect is interpreted as a measure of the strength of associative connections between the target concepts and the attribute values.

Problems with the IAT

Since the IAT effect is a relative measure, it cannot be precisely determined how it comes about. Different patterns of associations can in principle lead to the same IAT effect. Mierke & Klauer (2003) also found indications that the IAT effect can also occur without the existence of associations: "If associative links are only sufficient but not necessary for the effects to occur, an IAT effect cannot occur the existence of certain associations can be concluded ".

Another problem is the inadequate retest reliability , even if the internal consistency is partially sufficient. This is to be seen as problematic because semantic networks would have to be at least relatively stable over a certain period of time, even if they can basically be changed. Therefore, differential statements (i.e. about personality differences) should be treated with caution. The different test results for one and the same person are explained by the multi-dimensionality of the social identity , the self-stereotyping and the self-categorization of this person. Because of this, different strengths of association can be willingly changed or influenced by activating different self-concepts.

Furthermore, the basic assumption of the IAT, namely that an association between cognitive concepts is the basis for the variance of the reaction time, can be questioned. This is because figure-ground asymmetries and the resulting differences in salience can explain the results of the IAT equally.

The relationship between explicit and implicit association has so far been little researched.

IAT variants

Single-target IAT

In order to avoid the relativity of the IAT effect and to obtain an absolute measure of the evaluation of individual categories, only one target concept is presented per phase for the single-target IAT (Wigboldus, 2003). As with the classic IAT, people are presented with positive and negative words in the first phase, which they should classify as positive or negative by pressing a key. In contrast to the IAT, only one answer button is assigned twice in the subsequent phases. In the case of the gender IAT, in the second phase, people are asked to respond to positive words with the left button and to negative words and men's names with the right button. In the third phase, the double assignment of the keys is reversed: People are asked to respond to positive words and male names with the left key and to respond to negative words with the right key. Both phases are repeated, with the difference that women's names are used instead of men.

Extrinsic Affective Simon Task (EAST)

Example of a gender EAST. The colored words at the bottom of the screen represent the assignment of the reaction buttons.

The extrinsic-affective Simon problem (De Houwer, 2003) is based on a combination of IAT elements with the affective Simon effect . As with the IAT, two target concepts are used. The task of the person is to react to colored stimuli (mostly words) with one of two keys, each of which has a double function (stimulus color and valence). For example, people are asked to respond to positive white words with the left button and negative white words with the right button. In addition, they should respond to, for example, blue words (which correspond to one target concept in terms of content) with the left answer button, and to e.g. yellow words (which correspond to the other target concept in terms of content) with the right answer button. People usually classify the stimuli more quickly if the valence, which is also activated by processing the stimulus content, and the stimulus color are on the same answer key. In the example above, one would expect people to react faster to positive blue words than to positive yellow words, since positive and blue should be responded to with the same answer button.

Go / No-go Association Task (GNAT)

In the go / no-go association task (GNAT) by Nosek and Banaji (2001), in contrast to the classic IAT, only one target concept is used. This bypasses the problem of the relativity of the IAT results. It is the task of the person to react to certain stimuli (go), but not to others (no-go). The person is presented with positive, negative and stimuli of the target concept one after the other. In one round the person should react to positive stimuli and stimuli of the target concept, in another round to negative stimuli and stimuli of the target concept. The remaining stimuli (either negative or positive depending on the passage) should not be reacted to. As a rule, people react faster in the round that is compatible with their evaluation of the target concept than in the other round.

See also

literature

  • Banaji, MR, & Greenwald, AG (2015). Prejudices - how our behavior is unconsciously controlled and what we can do about it. Munich, dtv.
  • De Houwer, J. (2003). The extrinsic affective Simon task. Experimental Psychology, 50, 77-85.
  • Gawronski, B. & Conrey, FR (2004). The implicit association test as a measure of automatically activated associations: range and limits. Psychologische Rundschau, 55 (3), pp. 118–126.
  • Greenwald, AG, McGhee, DE & Schwartz, JLK (1998). Measuring individual differences in implicit cognition: The implicit association test (PDF; 513 kB). Journal of Personality and Social Psychology, 74, 1464-1480.
  • Kroeber-Riel, W., Weinberg, P. & Gröppel-Klein, A. (2008). Consumer behavior (9th edition) Munich: Vahlen.
  • Lane, KA, Banaji, MR, Nosek, BA & Greenwald, AG (2007). Understanding and using the Implicit Association Test: IV. What we know (so far). In B. Wittenbrink & NS Schwarz (Eds.). Implicit measures of attitudes: Procedures and controversies (pp. 59-102). New York: Guilford Press.
  • Nosek, BA & Banaji, MR (2001). The Go / No-go Association Task (PDF; 345 kB). Social Cognition, 19, 625-666.

Web links

Individual evidence

  1. ↑ a b Mierke, J. (2004) . Cognitive processes in the indirect measurement of attitudes with the Implicit Association Test. Dissertation to obtain a doctorate (PDF).
  2. ↑ Collins, AM , & Loftus, EF (1975). A spreading-activation theory of semantic processing. Psychological Review, 82, 407-428.
  3. ^ Meyer, DE & Schvaneveldt, RW (1971). Facilitation in recognizing pairs of words: Evidence of a dependence between retrieval operations. Journal of Experimental Psychology, 90, 227-234.
  4. ↑ Fazio, RH , Sanbonmatsu, DM, Powell, MC & Kardes, FR (1986). On the automatic activation of attitudes. Journal of Personality and Social Psychology, 50, 229-238.
  5. ↑ Mierke, J. & Klauer, KC (2003). Method-specific variance in the Implicit Association Test. Journal of Personality and Social Psychology, 85, 1180-1192.
  6. ↑ Cordelia Fine : The Gender Lie . Klett-Cotta, 2012, ISBN 978-3-608-10274-1 ( limited preview in Google book search). Part 1, Chapter 1
  7. ↑ Rothermund, K., & Wentura, D. (2004): Underlying processes in the Implicit Association Test: Dissociating salience from associations. Journal of Experimental Psychology: General , 133, 139-165. (PDF; 232 kB)
  8. ↑ Cordelia Fine : The Gender Lie . Klett-Cotta, 2012, ISBN 978-3-608-10274-1 ( limited preview in Google book search). P. 38
  9. ↑ Wigboldus, D. (2003). Single target implicit association tests. Paper presented at the 3rd Workshop on Implicit Diagnostics, Bonn, Germany.