What is crystallized intelligence what is fluid intelligence which type of intelligence increases with age?

Fluid Intelligence

Cattell's and Horn's fluid intelligence indexes people's ability to identify the underlying rules or concepts in novel problem-solving domains (Cattell, 1963; Horn, 1968). As Cattell stated, “Fluid general ability … shows more in tests requiring adaptation to new situations, where crystallized skills [domain-specific knowledge] are of no particular advantage” (Cattell, 1963, p. 3). Mathematics is an evolutionarily novel domain and thus fluid intelligence should be a significant contributor to individual differences in the ease of learning newly introduced mathematical concepts and particularly important as mathematics becomes increasingly abstract in later grades (Geary, 1995, 2005). In theory, however, the relative importance of fluid intelligence will decline once the concept is understood, but with the continuous introduction of new and more abstract concepts in the standard mathematics curriculum, fluid intelligence will remain important.

Most studies of the relation between intelligence and mathematics achievement have used a composite measure (e.g., standardized IQ test) that technically does not provide a direct assessment of fluid intelligence but will be highly correlated with it (Walberg, 1984). One of the largest of these studies included 70,000 students and found that intelligence measured at age 11 years was highly correlated (r = .77) with standardized mathematics achievement scores 5 years later (Deary, Strand, Smith, & Fernandes, 2007). In an analysis of almost 5000 children and adolescents, Taub and colleagues separated the contributions of other factors, such as crystallized intelligence, from fluid abilities and found the latter predicts mathematics achievement, albeit the strength of the relation was somewhat lower than that found by Deary et al. (β = .37 to .75) and varied across age (Taub, Keith, Floyd, & McGrew, 2008). Similar to the pattern shown in Fig. 1, contributions of crystallized knowledge and fluid intelligence to the prediction of adolescents’ mathematics achievement were about the same, but crystallized knowledge was not important for younger children. Although their measure of crystallized knowledge included many areas, not just mathematics, it should be a reasonable proxy for mathematics knowledge.

This is not to say that domain-specific knowledge is not particularly important for children's early mathematics learning; it depends on the mathematics being assessed. Indeed, a similar overall pattern emerges for the age at which preschool children learn their first mathematical concept—the cardinal value or quantities represented by number words—and for their mathematics achievement generally; a combination of prerequisite domain-specific knowledge (e.g., knowing the list of count words) and intelligence is important (Geary & vanMarle, 2016; Geary, vanMarle, Chu, Hoard, & Nugent, in press). The age at which children learned this mathematical concept and again fluid intelligence independently predict these children's school-entry number knowledge and mathematics achievement 3 years later (Geary, vanMarle, Chu, Hoard, et al., in press, Geary, vanMarle, Chu, Rouder, et al., 2018).

The pattern across all of these studies is consistent with Cattell's (1987) investment theory; “… this year's crystallized ability level is a function of last year's fluid ability level—and last year's interest in school work” (Cattell, 1987, p. 139). In other words, strong fluid abilities accelerate the learning of mathematics, assuming sufficient motivation and engagement with the material, and this domain-specific foundation along with fluid abilities jointly contribute to future learning. Environmental factors such as the quantitative activities at home and quality of curricula material will be important for exposing children to the appropriate mathematics content (LeFevre et al., 2009; Ramani & Siegler, 2008).

Generic learning outcomes

6.2.2 Fluid (or Performance) Intelligence

As noted above, fluid intelligence is the aspect of intelligence that is minimally dependent on language. One research team described fluid intelligence as “our ability to creatively and flexibly grapple with the world in ways that do not explicitly rely on prior learning or knowledge” (Tranter & Koutstaal 2008:185). Fluid intelligence allows individuals to often intuit the solutions to puzzles and problems, particularly those of a spatial or mathematical nature, often with little use of language (Schretlen et al. 2000:55).

Studies have shown that fluid (or performance) intelligence declines with age, normally starting in one’s early to mid-20s (Kaufman & Horn 1996; Isingrini & Vazou 1997; Tranter & Koutstaal 2008). Fortunately, an individual’s crystallized intelligence is able to at least partially compensate by continuing to grow throughout most of one’s life (Cunningham et al. 1975; Kaufman & Horn 1996).

As one can see by inspecting Table 6.2.2, all but one of the studies of fluid IQ has concluded that it is positively correlated with social status. The exception was a German study which found a positive correlation for males but no significant correlation for females.

Table 6.2.2. Relationship Between Social Status and Fluid (or Performance) Intelligence

PIQ, performance IQ.

Abstract

This chapter explores the interplay between fluid intelligence declines and higher levels of crystallized intelligence of older adults as they affect everyday decision-making ability. Specifically, we explore the hypothesis that accumulated knowledge and expertise may help compensate for age-related declines in fluid cognitive function. The complementary capabilities framework suggests that although age-related declines are inevitable, these declines may be at least partially attenuated on tasks and in domains that are more familiar and practiced. Crystallized intelligence may thereby represent a kind of intellectual capital that circumvents reduced capabilities caused by diminished levels of fluid intelligence. We explore the role of domain-specific knowledge and expertise in context-specific tasks and everyday problem solving, and discuss the practical implications of this research for public policy and for the design of effective decision interventions that can aid decision making among older adults.

Fluid Intelligence

Most studies of abstract reasoning and fluid intelligence have demonstrated a significant relation between WMH burden and these abilities. Longitudinal increase in WMH volume over five years was associated with longitudinal decline in fluid ability (Raz et al., 2007). Similarly, Garde et al. (2000) found in a cohort of 80 year olds who had been tested with the WAIS multiple times throughout their lives that both periventricular and deep WMH burden was associated with longitudinal decline in intelligence, primarily on tests that measure performance rather than verbal abilities, perhaps due to the motor and/or speeded nature of those tasks. In line with these findings, Vannorsdall et al. (2009) found periventricular WMH burden to be negatively associated with fluid, but not crystallized, intelligence scores. Cook et al. (2002) also found a significant association between lower abstract reasoning and increased WMH burden.

Fluid and crystallized intelligence

Another way of splitting up intelligence is into fluid intelligence as against crystallized intelligence and as we shall see, this is a very important division in relation to age effects.

Fluid intelligence is composed of the ability to be flexible and to respond adaptively to novel situations. The exercise of fluid intelligence requires effortful processing and attention. It would be identified with Type 2/System two thinking (as discussed in Chapter 2) and tasks tapping fluid intelligence would load working memory. A real-life example of using fluid intelligence would be in solving new abstract problems as in choosing between different complex savings schemes or working a new piece of electronic equipment and in creative thinking generally.

Crystallized intelligence is an ability to use knowledge acquired through life experience, and it draws on information, language, and established skills. The exercise of crystallized intelligence is often relatively automatic. It can be identified with Type1/System one thinking and tasks tapping crystallized intelligence would not load working memory. Examples of real life use of crystallized intelligence would be in remembering needed facts, such as what fertilizer is suitable for which plants, and in carrying out well-practiced, sensory-motor skills such as driving a car.

2.1 Differential Age-gradients of Cognitive Mechanics and Pragmatics

Traditionally, two-component models of intelligence distinguish between fluid intelligence reflecting the operations of neurobiological ‘hardware’ supporting basic information-processing cognitive mechanics and crystallized intelligence reflecting the culture-based ‘software’ constituting the experience-dependent cognitive pragmatics (Baltes et al. 1999, Horn 1982; see also Lifespan Theories of Cognitive Development). Figure 3 shows that the fluid mechanics such as reasoning, spatial orientation, perceptual speed, and verbal memory show gradual age-related declines starting at about the 40s, while other abilities indicating the crystallized pragmatics such as number and verbal abilities remain relatively stable up until the 60s (e.g., Schaie and Willis, 1993). Furthermore, there have also been some recent ongoing theoretical and empirical efforts devoted towards expending the concepts of cognitive mechanics and pragmatics. In addition to the efficacy of information processing, cognitive mechanics also encompasses the optimal allocation of cognitive resources (e.g., Li et al. in press). Cognitive pragmatics has been expanded to include many other general as well as person-specific bodies of knowledge and expertise associated with the occupational, leisure, and cultural dimensions of life (e.g., Blanchard-Fields and Hess 1996). One example is wisdom, the ‘expert knowledge about the world and fundamental pragmatics of life and human affair’ that an individual acquires through his or her life history, that also includes an implicit orientation towards maximizing individual and collective well-being (Baltes and Staudinger 2000).

3.4.2 The n-Back in Action

The positive influence of dual n-back training on fluid intelligence reported in Jaeggi’s first study (Jaeggi et al., 2008) was also observed in a group of undergraduates from the National Taiwan Normal University in Taipei (Jaeggi et al., 2010). In this study, two groups of students were trained for 4 weeks on either a single or a dual n-back task. Both groups, in comparison with a no-contact control group, showed significant improvements on two tests of fluid intelligence. An interesting finding was that single n-back training was equally effective as dual n-back training.

Similar improvements were also observed in a group of elementary and middle school children (Jaeggi et al., 2011). In this study, children in the experimental group trained on an adaptive spatial single n-back instead of a dual n-back task. By contrast, children in the active control group learned how to answer questions related to general knowledge and vocabulary. This intervention was assumed to promote skills related to crystallized intelligence. Effects on trained tasks were only observed for the experimental group; however, there were no transfer effects on measures of fluid intelligence. In an additional analysis, the experimental group was split in relation to training performance (below/above average training gains). Far transfer effects on fluid intelligence were only present for students showing pronounced training gains on the n-back, whereas no effects were found for the control group, and the students in the experimental group that displayed low n-back training gains. Furthermore, far transfer effects remained stable even after a 3-month hiatus from training. The authors concluded that far transfer effects can only be expected in the presence of training gains on the trained tasks. A questionnaire revealed that although the training was adaptive, some children who had improved the least reported that they found the tasks difficult and effortful.

To further analyze training characteristics that may improve far transfer effects Jaeggi et al. (Katz et al., 2014) tested seven versions of a game-like n-back training paradigm. All training versions were adaptive and designed to measure the influence of five motivational features: points (a bar at the bottom of the screen displayed the actual score); theme changes (several different themes were included using different characters, such as a frog or a cat); lives and levels (a level indicating n was shown, as well as lives, indicating the number of errors that could be made before dropping to a lower n level); prizes were offered after training in exchange for points, and end-of-session certificates were awarded showing the level of n a person reached. A surprising finding was that none of the provided incentives affected training improvement. On the contrary, children performed better when no points, levels, or lives were shown compared to a group of children that were exposed to all five motivational features. These findings are counterintuitive and hard to explain. The authors suggested that given the demanding nature of the n-back task, additional information might have distracted the attention of trainees.

In yet another study, Jaeggi et al. (2014) investigated whether individual differences of trainees (their need for cognition and their implicit theories about intelligence) influence transfer training effects. Students were randomly assigned to three training groups: a dual n-back; a single n-back where subjects trained on just the auditory part of the dual n-back task, and an active control group where subjects learned how to answer questions related to general knowledge and vocabulary. The results replicated previous findings indicating that both training groups showed far transfer effects to a composite score representing five visuospatial reasoning measures. Of particular interest is the finding that training on the n-back task presented in the auditory stream resulted in transfer to measures of visuospatial reasoning, which implies that the transfer is modality independent. The moderator variable need for cognition, which indicated how much one enjoys cognitively challenging tasks, did not significantly influence transfer effects. On the other hand, individuals who thought that intelligence can be changed (is malleable) showed more transfer to visuospatial ability. Besides replicating previous findings, this study is important because it resolved several methodological shortcomings that were present in previous studies conducted by the same research group.

Another question that attracted the researchers that had used the n-back as a training task was what happens in the brain when training improves performance on a trained task? To answer this question Buschkuehl et al. (2014) used a technique called arterial spin labeling (ASL) which is from a methodological viewpoint more suited for longitudinal studies (such as cognitive training studies) than the more often used BOLD measure that is prone to scanner drifts (e.g., Hernandez-Garcia and Buschkuehl, 2012). The participants in the experimental group trained on an adaptive visuospatial n-back task whereas the control group answered vocabulary and general knowledge questions. Before and after training, the participants solved an easy 1-back and a more difficult 4-back task in a MRI scanner. The MRI analysis used perfusion as a surrogate for neural activity indicating increases in activity (contrast difference 4-back −1-back) in several brain areas. The results showed that higher task proficiency was related to increased resting perfusion in frontal brain regions accompanied with decreased activity in one parietal cluster. A combined analysis of all factors (e.g., rest vs. n-back proficiency and control vs. training) revealed increased activity after the intervention in the BA6 area located in the left precentral gyrus/frontal middle gyrus/superior frontal gyrus, a finding that is in line with previous research (Section 3.3.1). The authors concluded that the MRI analysis revealed that training on the n-back increased the physical fitness of the brain—a condition that might also explain other transfer effects.

8.5.1 Experiment Design

In this study, the participants were divided into two groups: one group of participants was tested with 2D learning contents while another group was tested with 3D learning contents. As mentioned in Chapter 7, 2D and 3D Educational Contents, the individual participants’ reasoning and general cognitive ability were assessed using Raven’s Advanced Progressive Matrices (RAPM) test. The RAPM task was also required while analyzing the ERP components for mental fatigue. The reason is that the relationship of ERP components and general intelligence is reported in previous studies. Hence, the RAPM data of the participants was also required in this chapter along with the learning task (watching learning animated contents with 3D or 2D display) and oddball tasks for ERP signal analysis. The RAPM and learning task is briefly touched on here, and the oddball task is described in detail.

The following is the list of the experimental tasks used in the study:

Raven’s Advanced Progressive Matrices (RAPM) test

3D visualization

oddball task

8.5.1.3 Visual Oddball Task

The oddball paradigm is a commonly used task for cognitive and attention measurement in ERP studies.38–40 In this study, two visual stimuli, a box and a sphere, shapes of size 5 cm, were designed as the standard and target stimuli, respectively (see Fig. 8.2). The presentation duration of each trial, either the standard (box) or target (sphere) trial, was 500 ms with the intertrial interval (ITI) between two consecutive trials being 500 ms. The participants were instructed to press “0” for a target stimulus and not to respond for a standard stimulus. Further, the reaction time and correct target detection of each participant were recorded. Two types of error were expected: false alarm (i.e., pressed key when standard stimulus was shown) and omission (forgot to press key when target stimulus appeared). Thirty percent of the trials were kept target and 70% were nontarget trials, i.e., there were 40 target trials and 135 total trials presented. Total time spent on the oddball task was 3.35 minutes.

Figure 8.2. Visual stimuli of oddball task36 (box represents the standard stimulus and sphere represents the target stimulus).

2 Training of Basic Processes vs. Training of Everyday Activities

The distinction between two components of human intelligence, i.e., fluid intelligence as an age-related ability to solve new and unfamiliar problems and crystallized intelligence as an ability to solve familiar problems that can be preserved or even improved in old age (Horn 1982), does not mean that these components are independent of each other. Since every complex cognitive activity contains elements from fluid and crystallized intelligence and intellectual performance as a product can result from different proportions of the two components, expertise, i.e., a high level of crystallized intelligence, offers opportunities to compensate for losses in fluid intelligence.

The possibility to compensate for losses in basic cognitive processes has been proven in numerous empirical studies, especially in the field of occupational activities, but also in other meaningful everyday activities. It has been shown that performance in complex cognitive tasks does not decrease as fast as could be supposed from decreases in basic cognitive processes (Willis 1987). Strategies that allow for compensation in basic cognitive processes are, e.g., an intentional slowing of action, additional checks of solutions, restriction to a small number of activities and aims. However, as could be shown in the testing-the-limits paradigm, compensation in favor of the optimization of specific aspects generally leads to a prolongation of the time required for the task (Baltes and Baltes 1990, Kliegl et al. 1989).

The proven possibility to compensate for losses in intellectual abilities leads to the question whether everyday competence in old age can be improved by training of useful strategies and basic processes. In this context the person-centered intervention approach of Willis (1987) is instructive. According to this author complex everyday activities can be optimized by a training of basic processes. In the first step the significance of specific processes for clusters of important daily activities (e.g., reading operating instructions or an instruction leaflet) has to be determined. In a second step those processes that have an impact on performance in numerous activities can be trained. A training of basic processes would be very attractive for intervention research since participation in training programs could heighten performance in numerous contexts and activities. However, basic cognitive processes are at the very beginning of everyday performance; the relationship between the two is only poor and a satisfying prognosis of performance from basic processes is not possible. As a consequence, recent development in intervention research indicates a preference for another paradigm: the training of specific everyday activities. Since the context-independent training mnemonics failed to have the expected impact on everyday memory performance, it was proposed to offer specific courses aimed to improve memory of names or prevent people from mislaying glasses or keys instead of courses aimed to improve general memory performance. Following this approach it is necessary to create contexts of person-centered intervention that correspond very much to problematic situations in everyday living.

Consequently, from the perspective of this approach a detailed examination of individual life situations is demanded. This demand illustrates the principal dilemma of person-centered intervention programs: the expenditure of training so many people in so many specific situations is out of all proportion to the possible intervention effects. Intervention programs are often used to search for potentials for action and development, especially in the age-related component of intelligence. Numerous empirical studies have differentiated our understanding of human intelligence by demonstrating reserves of capacity for intellectual performance. Cognitive functions can be improved through adequate training programs, especially when individual, social, and occupational aspects of the life situation are taken into account. Moreover, cognitive training can also be helpful for reaching noncognitive aims, another indication of the significance of cognition for successful management of life in our culture.

However, effects of cognitive training remain specific for concrete problems and situations. Moreover, according to Denney (1994), most training studies (naturally) focus on age-related abilities and skills where similar gains can be reached through exercise alone. Additionally, training has the greatest impact on skills that are not needed in everyday life. Therefore, Denney (1994) raises the question why people should participate in conventional training programs and whether it would not be better to create new programs that concentrate on well-developed abilities and skills, where little effects could have a great impact on possibilities to maintain an independent and self-responsible life.