Intelligence theory

from Wikipedia, the free encyclopedia

There are various theories of intelligence (also called intelligence models) in differential psychology that attempt to describe the causes and effects of intelligence .

Factorial theories

Factorial theories about human intelligence are based on the method of factor analysis . This procedure enables the results of a large number of test items to be combined into a data structure that can be described by a few factors.

Spearman's Two Factor Theory

Charles Spearman discovered positive correlations between different intellectual achievements; if a person had above-average values ​​for intellectual achievement (e.g. as a school grade in German), it was more likely that further above-average achievements (e.g. in mathematics) could be recorded. Spearman defined this fact as a positive manifold . Observation of these positive (if not entirely perfect) correlations of tasks requiring intellectual ability formed the basis for Spearman's further research. According to Spearman, the intercorrelation of different intellectual achievements would have to be created through a common dimension; by a factor of general intelligence g (g = general). G should therefore represent a common basis or a universal connection across different tasks and situations.

Spearman developed his two-factor theory using the general factor model of factor analysis. A single superordinate factor is extracted from the available test data. In this way Spearman extracted the factor g . He described this factor as a general intelligence factor that influences all different performance areas. The expression of this general intelligence determines u. a. Processing speed, mental capacity, intellectual performance - in short: whether a person is more of a “simple character” or a “more talented genius”.

Since the correlation of the various intellectual achievements with g remained only medium-high even after a measurement error had been corrected, Spearman suggested that there must be another phenomenon that stands for each individual achievement. And so Spearman established the specific factor s (s = specific). If both factors are now taken into account, the equation for measuring a power measured value would look like this: x i = g i + s i

These specific intelligence factors are hierarchically subordinate to factor g and area-specific, independent factors. They determine (but significantly influenced by the factor g ) a person's performance in certain areas - e.g. B. in mathematical tasks, verbal or spatial problems.

The more pronounced the talent was in a certain intellectual ability, the more the factor g lost importance in favor of the specific factors. Spearman based this observation on the talent differentiation hypothesis. In addition, the factor g proved to be very robust in further studies of the same data set, despite different operationalizations and methods. Spearman defined this fact as the indifference of the indicators .

According to Spearman, the factor g is characterized by three basic operations:

  • Grasping the experience
  • Discovering / deriving relationships
  • Discovering / deriving connections

Since the factor s exists for every individual performance (s 1 , s 2 , s 3 , ..., s n ), the name "two-factor theory" is, strictly speaking, misleading. In addition to the general main factor g , any number of factors s characterize this model.

Thurstone's primary factor model

Louis Leon Thurstone rejected the idea of ​​a general, superordinate intelligence factor. He emphasized the area-specific organization of intelligence and viewed this as a combination of various individual skills. Through his factor analytical approach, he extracted seven primary mental abilities of intelligence:

  • S (space): spatial-visual tasks such as B. Mental rotation of objects
  • P (perceptual speed): perception of objects and relation between them, e.g. B. Continuation of a series of objects
  • N (numerical ability): numerical and mathematical skills
  • M (memory): memory performance, e.g. B. Answer questions about a scene that is shown for a short time
  • R (reasoning): logical reasoning
  • W (word fluency): verbal fluency, e.g. B. Finding synonyms
  • V (verbal relations): Understand and interpret verbal relationships correctly

One criticism often made against Thurstone relates to his methodical approach: to extract his factors, he uses what is known as an oblique transformation. The consequence of this is that the extracted factors are not completely independent of one another - that is, they correlate weakly with one another. In fact, there is a weak positive correlation between the seven primary factors. According to their theory, representatives of the general factor model of intelligence see the reason for this correlation in a superordinate, general intelligence factor (for the methodical interpretation of the correlation see partial correlation ).

Cattell's two-factor model

Even Raymond Cattell stood up against a hierarchical model concept. He identified two intelligence factors, fluid (or liquid ) and crystalline (also crystallized ) intelligence. The fluid intelligence is innate or inherited and can not be influenced by the environment. It includes, for example, the mental capacity, the comprehension, the general processing level.

The crystalline intelligence (also: crystallized intelligence ) comprises all skills that are learned in the course of life or determined by the environment. The crystalline intelligence is dependent on the fluid intelligence. It includes both explicit knowledge (semantic and episodic, such as factual knowledge) as well as implicitly learned knowledge (certain behaviors, cycling, arithmetic, etc.). Intelligence refers to the ability to apply this acquired knowledge .

In 1963, Cattell presented his model of "fluid and crystallized general intelligence", where he also took up and modified Spearman's model. He also carried out several factor analyzes and came up with three levels of order. The higher the order, the more general the factors are. There are six first-order factors, namely verbal, spatial, logical and numerical skills , as well as fluency and memory . The 2nd order factors are then divided into fluid and crystalline intelligence, to which the greatest attention is paid. Finally, the 3rd order factors are historical fluid intelligence and general learning experience. The two 2nd order factors have different properties. Fluid intelligence is responsible for analyzing tasks. Above all, it includes innate capabilities and should therefore be viewed as more general and instinctive . The mere ability, the capacity to acquire knowledge is also justified by the fluid intelligence. Skills such as logical thinking or the creation and use of complex relationships are subordinate to this factor and above all determine the ability to adapt to new problems and situations. Culture-free tests can be used to make this part of intelligence measurable ; this means that these tests do not relate to general knowledge , which is culturally different. One can assume that the fluid intelligence is influenced by the respective test situation. In general, it can be said that fluid intelligence is very much tied to intact neural structures and processes and can accordingly be impaired by illness or injury. The development comes to a standstill around the age of 14/15, and from the age of 22 it even declines somewhat.

The fluid intelligence z. Sometimes dependent crystalline intelligence (cf. investment theory ), on the other hand, relates to the execution of a job, the solving of a task, specifically in relation to education , knowledge. There are now the mentioned culture-specific elements. The stored knowledge, the previous learning processes come to the fore here. The factor is particularly evident in verbal, numerical or mechanical skills in the mind and judgment . Crystalline intelligence is best assessed with culture-specific tests. Because the crystalline intelligence contains the knowledge of a person, an easy connection to the personality can be established. It is strongly influenced by practice and interest. In the case of crystalline intelligence, development is largely over between the ages of 18 and 20, but it can also extend to the age of 50. Cattell noted in 1973: "Crystalline intelligence is in a sense the end product of what fluid intelligence and education have brought about together."

What should be mentioned for understanding is that the crystalline intelligence is not to be equated with the achievement; because it relates to dealing with complex contexts, while performance covers all academic knowledge of the individual. It can thus be said that tests carried out so far lose their validity if the model is accepted, since crystalline and fluid intelligence were never recorded separately from one another.

The CHC model

In his analysis of hundreds of studies on intelligence, John B. Carroll (1993) was able to show that many of the factor models mentioned here can be integrated. His investigation resulted in a multilevel hierarchical model. At the lowest level (Stratum I) there are highly specific tasks. The middle level (Stratum II) contains more complex skills, for example fluid intelligence Gf or crystalline intelligence Gc . At the highest level (Stratum III) is the general cognitive ability g , with which all subordinate properties are more or less strongly related.

Since Cattell's model of fluid and crystalline intelligence (e.g. Horn & Cattell, 1966), which is particularly important in the context of intelligence development , can be subsumed under Carroll's results, McGrew (2005) proposed that these two be combined into one model (“CHC model”), which has since enjoyed wide acceptance in research, but has also been criticized for its vague conceptualization. One of the main criticisms of the CHC model relates to the idea on which the crystalline intelligence Gc is based that it arises due to individual differences in the fluid intelligence Gf . In this conceptualization, Gc can be understood as a formative construct that should be able to be completely conditioned to Gf . Since this was not the case in Carroll's analysis, only unobserved third-party variables can be responsible for the dissociation of Gf and Gc. In a reanalysis of the data set used by Carroll, Kan and colleagues were able to show that, taking into account different educational backgrounds, Gf on Stratum II and g on Stratum III are completely identical. Gc, on the other hand, disappeared from the factor solution. Against this background, crystalline intelligence should not be understood as an independent (organic) intelligence facet, but can be explained by the environmental, individually different access to education and the associated training of verbal skills.

Multi-dimensional models

Guilford's cube model

A factorial approach to the study of intelligence comes from Joy Paul Guilford . This differentiates between three dimensions of intelligence.

  • On the one hand, the thinking content . Here he distinguishes four classifications - for example abstract or figural thought content.
  • The second dimension is represented by the thought operations . These are divided into five levels, e.g. B. convergent approach (the concentration on a specific solution approach and the consistent further development of this) or divergence approach (finding as many different possible solutions as possible and ultimately choosing the best).
  • The third dimension is determined by the thinking outcomes . These are divided into six categories, e.g. B. Finding a new, unique solution, finding categories or classes, or transferring the solution from one situation to another.

These three dimensions, represented graphically, span a three-dimensional coordinate system. In this one can now imagine a cuboid, on the three visible surfaces of which all possible combinations of the three dimensions are shown in small rectangles (therefore a tetrahedral model, since these three surfaces of the cuboid are important). According to Guilford, the (4 × 5 × 6 =) 120 combinations each represent individual areas of intelligence. Representatives of this approach have not yet fully succeeded in finding suitable tasks for each combination (around 20 of them are still pending).

Guilford's tetrahedral model is sometimes referred to as "Guilford's intelligence structure model" or "ISM by JP Guilford". In current intelligence research, the model only plays a role from a historical perspective.

Jäger's Berlin Intelligence Structure Model (BIS)

Berlin intelligence structure model after Adolf Otto Jäger

Another approach is the Berlin intelligence structure model by Adolf Otto Jäger (1984). In his research project "Productive Thinking / Intelligent Processing", he tried to confront the competing models with the most representative sample of variables of the performance area for which validity is claimed and to create a structural model based on sample variables that represent the diversity of intellectual performance forms as comprehensively as possible to generate. The representative variable sample was ensured by extracting 191 task blocks from approx. 2000 task types, which in turn could be assigned to 98 task types. This means that the diversity of the task material of the various models has been retained. The test subjects were Berlin high school students aged 16 to 21 years.

In Jäger's work, a descriptive model was created, which is structured hierarchically and bimodally . Jäger extracts seven highly general main components in two established modalities, whereby these identify different aspects under which the same objects can be classified. The "Operations" modality consists of the following components:

  • Processing speed (B; pace of work, ease of perception and power of concentration when solving simply structured tasks with a low level of difficulty),
  • Retention (M; active memorization and short- or medium-term recognition or reproduction of verbal, numerical and figurative-pictorial material),
  • Ingenuity (E; fluid, flexible and also original production of ideas, which requires the availability of diverse information, wealth of ideas and seeing many different sides, variants, reasons for possibilities of objects and problems, whereby it is about problem-oriented solutions, not uncontrolled indulgence in fantasies) and the
  • Processing capacity (K; processing of complex information in tasks that require a variety of relationships, formal-logical exact thinking and appropriate assessment of information).

The "Content" modality consists of the skill bundles:

  • language-bound thinking (V; verbal; bundle of skills corresponds to the degree of its acquisition and availability and seems to be a determining factor in all language-related operations),
  • number-based thinking (N; numerical; bundle of skills corresponds to the degree of its acquisition and availability and seems to be involved in all number-based operations) and the
  • perception-bound thinking (F; figural-pictorial).

The general intelligence “g” includes all seven of the main components mentioned. Jäger does not consider the structural components listed as well as "g" to be final; they should rather be viewed as a model core that is open to the addition of further operational and content-bound units, the settlement of units between “g” and the seven main components, differentiations into more specific units and the addition of further modalities. For example, further investigations in the research project “Productive Thinking / Intelligent Processing” on the subject of practical intelligence could show that it would be appropriate to expand the Berlin intelligence structure model to include a further content modality, “concrete, objective material”. Current research findings on the subject of auditory intelligence also indicate that a model extension to include a content modality "auditory" appears appropriate.

The Berlin intelligence structure model was able to be replicated in various international samples, with different tasks and different evaluation methods. Another special feature of the BIS model is that the BIS test is a content-valid intelligence test that corresponds to the model and covers all facets and dimensions of the BIS model (in contrast to the CHC theory, for example ).

The Radex model

Guttman (1954) postulated that intelligence test tasks differ in terms of their complexity. With the aid of the multidimensional scaling process , this relationship can be graphically illustrated. More complex types of tasks are closer to the center of Radex. Guttman further distinguished between three content areas (figural, verbal and numerical intelligence), which are arranged in the form of sectors around g. Mathematical word problems should therefore be located in the area of ​​overlap between the numerical and the verbal sector. This differentiation between different content areas was confirmed in the Berlin intelligence structure model (BIS).

Marshalek et al. Suggested an integrative approach between the hierarchical and the Radex approach. (1983). They were able to prove that the general intelligence (g), which is higher in the hierarchy, is to be equated with Guttman's complexity. Tasks with a high g-charge (such as Raven's matrix tasks) are accordingly close to the center.

Basic cognitive processes: processing speed and working memory

Compared to more complex skills such as logical reasoning, processing speed and working memory represent more fundamental mental skills, the importance of which for intelligence has been discussed since the beginning of intelligence research and has not lost its relevance to this day. Siegfried Lehrl, for example, advocates the theory that intelligence is based on information processing speed and memory span . The working memory is located in the prefrontal brain. He also designed the short test for general basic quantities in information processing .

Processing speed

Francis Galton (1883) took the view early on that various measures of cognitive speed such as reaction time allow conclusions to be drawn about a person's intelligence. More recent empirical findings on the development of intelligence over the life span indicate that a decrease in mental speed is accompanied by a decrease in fluid intelligence (in the sense of Cattell), while crystalline intelligence remains unaffected (Finkel et al., 2007) and thus confirm the importance of the speed component. However, the question has not yet been finally resolved; A further complicating factor is that mental speed is also a significantly more complex construct than previously assumed (cf. Nettelbeck, 2011).

Working memory

Kyllonen and Christal (1990) postulated that working memory and reasoning (as the central component of intelligence; see Guttman's Radex model) are essentially the same construct. This was confirmed by the findings of Suss et al. (2002), who specifically examined working memory capacity. A meta-analysis by Ackerman et al. (2005) refuted the assumption that intelligence and working memory are identical; the correlations adjusted for measurement errors are only in the middle range around r = .48. However, these findings also did not go unchallenged (e.g. Oberauer et al., 2005; Kane et al., 2005).

Information Processing Theory

Information processing theory rejects the idea of ​​fundamental factors of intelligence. Rather, it deals with the cognitive processes that take place during information processing. Essentially three questions are of interest here:

  • What kind of cognitive process is going on?
  • How accurately is this process carried out (i.e. how fast, how complex, etc.)?
  • What mental representation is this process based on (i.e. is someone thinking in pictures, or in abstract numbers, etc.)?

Sternberg's triarchic model (component model)

One of the most important representatives of the information processing approach is Robert Sternberg . In his triarchic model he postulates three theories:

  • Context theory: Everyone has a culture-specific or environment-specific intelligence. This enables him to integrate himself into his environment, to establish and maintain social contacts and to follow cultural norms more or less.
  • Two-facet theory: In order to study intelligence, it is not only necessary to record the underlying solution processes, the implementation and results. In addition, it is important to record the routine or automation of the processes, as this has an important influence on the accuracy and result of a solution strategy.
  • Component theory: Sternberg distinguishes five components of cognitive processes
1) Performance components: These are area-specific skills or solution strategies. A calculation task requires z. B. an abstract mathematical solution strategy, a word task, however, more verbal skills.
2) Metacomponent: This largely corresponds to a higher-level executive control. It decides which performance components are used in a particular situation.
3) Acquisition component: Reference is made here to the storage or encoding of information. Like 4), this component corresponds to a memory function.
4) Retention component: This refers to the retention and retrieval of information from memory.
5) Transfer component: The last component concerns the transfer of knowledge or skills learned in a certain situation to other problems and situations.

Another merit of Sternberg lies in his expansion of the concept of intelligence. Intelligence therefore comprises learning from experience , abstract reasoning , the ability to adapt to a constantly evolving and changing environment and the motivation to acquire new knowledge or skills . The first two points are already covered by popular intelligence tests. However, the last two points have so far been given little or no consideration when recording “intelligence”.

Multiple intelligence according to Gardner

Howard Gardner takes the view that we do not have one, but several independent intelligences - that is, a multiple intelligence . This theory is called the theory of multiple intelligences . He not only goes so far as to subdivide these intelligences into area-specific units (similar to some factor theories), but also locates them in independent module-like organizational forms in the brain. Every intelligence should be based on its own neural "circuit" in the brain. Impairments or injuries to one intelligence should therefore not have any influence on other intelligences.

Another classification of Gardner's intelligence includes two areas: Intrapersonal intelligence refers to the knowledge about oneself, the interpretation of one's own feelings and behavior, the prediction of one's own behavior, etc. Interpersonal intelligence comprises interpersonal parts of knowledge and skills. For example, predicting the behavior of other people, empathic skills, the ability to behave according to the expectations of others, etc.

However, many intelligence researchers criticize Gardner's work as not being supported by scientific research results.

Jensen and Eysenck: Two basic processes of intelligence

Arthur Jensen and Hans Jürgen Eysenck assume that there are two basic processes of intelligence. They refer to these as Level I abilities and Level II abilities (in German: Level I skills and Level II skills). Level I (associative ability) receives the neural registration and consolidation of the stimulus inputs and the formation of associations. Level II (cognitive, conceptual skills) includes the evaluation of the stimuli. Conceptual learning and problem solving are good examples. Intelligence tests , especially language tests, are used to assess level II skills. The so-called number connection test ( Trail making test ; numbers or letters that are randomly distributed over a sheet of paper should be connected with a line in the correct order) is suitable for assessing the skills of level I. There is little correlation between Level I and Level II skills.

See also

Individual evidence

  1. Suess, H.-M. (2003): Intelligence Theories. In K. Kubinger, & RS Jäger (ed.), Keywords of Psychological Diagnostics. (Pp. 217-224). Weinheim: Psychology Publishing Union. ISBN 978-3621274722 .
  2. Intelligence, crystallized and fluid in DORSCH Lexikon der Psychologie, ISBN 978-3-456-85643-8
  3. Carroll, JB (1993). Human cognitive abilities: A survey of factor-analytic studies. Cambridge: Cambridge University Press.
  4. Horn, JL & Cattell, RB (1966). Refinement and test of the theory of fluid and crystallized general intelligences. Journal of Educational Psychology, 57, 253-270.
  5. McGrew, KS (2005). The Cattell-Horn-Carroll theory of cognitive abilities. Past, present and future. In DP Flanagan & PL Harrison (Eds.), Contemporary intellectual assessment: Theories, tests, and issues (2nd edition, pp. 136-181). New York, NY: Guilford Press.
  6. Kan, Kievit, Dolan, & van der Maas (2011). On the interpretation of the CHC factor Gc. Intelligence, 39, 611.
  7. Riffert, F. (2010). Educational diagnostics - objective procedures - intelligence (p. 15). Salzburg: Paris Lodron University (Scriptum).
  8. ^ Stern, E., & Neubauer, A. (2016). Intelligence: not a myth, but reality. Psychological Rundschau, 67 (1), 15–27. doi : 10.1026 / 0033-3042 / a000290
  9. a b Süß, H.-M., & Beauducel, A. (2011). Intelligence tests and their relation to intelligence theories. In LF Hornke, M. Amelang, & M. Kersting (Eds.), Performance, Intelligence and Behavioral Diagnostics (Vol. 3, pp. 97-234). Göttingen: Hogrefe.
  10. Conzelmann, K., & Süß, H.-M. (2015). Auditory intelligence: Theoretical considerations and empirical findings. Learning and Individual Differences. doi : 10.1016 / j.lindif.2015.03.029
  11. Süß, H.-M., & Beauducel, A. (2015). Modeling the construct validity of the Berlin Intelligence Structure Model. Estudos de Psicologia (Campinas), 32 (1), 13-25. doi : 10.1590 / 0103-166X2015000100002
  12. ^ Guttman, L. (1954). A new approach to factor analysis: The radex. In PF Lazarsfeld (Ed.), Mathematical thinking in the social sciences (pp. 258-348). Glencoe, IL: Free Press.
  13. Jump up Marshalek, B., Lohman, DF, & Snow, RE (1983). The complexity continuum in the radex and hierarchical models of intelligence. Intelligence, 7, 107-127.
  14. ^ Galton, F. (1883). Inquiries into human faculty and its development. New York, NY: AMS Press.
  15. Finkel, D., Reynolds, CA, McArdle, JJ, & Pedersen, NL (2007). Age changes in processing speed as a leading indicator of cognitive aging. Psychology and Aging, 22, 558-568.
  16. ^ Nettelbeck, T. (2011). Basic processes of intelligence. In RJ Sternberg & SB Kaufman (Eds.), The Cambridge handbook of intelligence (pp. 371-393). New York, NY: Cambridge University Press.
  17. Kyllonen, PC & Christal, RE (1990). Reasoning ability is (little more than) working-memory capacity ?! Intelligence, 14, 389-433.
  18. ^ Suss, H.-M., Oberauer, K., Wittmann, WW, Wilhelm, O. & Schulze, R. (2002). Working memory capacity explains reasoning ability — and a little bit more. Intelligence, 30, 261-288.
  19. ^ Ackerman, PL, Beier, ME & Boyle, MO (2005). Working memory and intelligence: The same or different constructs? Psychological Bulletin, 131, 30-60.
  20. Oberauer, K., Schulze, R., Wilhelm, O. & Süß, H.-M. (2005). Working memory and intelligence — their correlation and their relation: Comment on Ackerman, Beuer, and Boyle (2005). Psychological Bulletin, 131, 61-65.
  21. Kane, MJ, Hambrick, DZ & Conway, ARA (2005). Working memory capacity and fluid intelligence are strongly related constructs: Comment on Ackerman, Beier, and Boyle (2005). Psychological Bulletin, 131, 66-71.
  22. Detlef H. Rost : Multiple intelligences, multiple irritations. Journal for Educational Psychology 22 (2008) 97–112
  23. H. Weber, H. Westmeyer: The inflation of the intelligences . In E. Stern & J. Guthke (Ed.): Perspektiven der Intellektivenforschung . Pabst, Lengerich 2001, pp. 251-266.
  24. Heinz-Martin Süß, André Beauducel: Intelligence tests and their references to intelligence theories . In LF Hornke, M. Amelang, & M. Kersting (eds.): Performance, intelligence and behavior diagnostics (Volume 3). Hogrefe, Göttingen 2011, pp. 97–234
  25. Eysenck, Hans Jürgen (1984): The inequality of people . Orion-Heimreiter-Verlag , Kiel. ISBN 3-89093-100-6 , p. 244

Web links