essay about individual differences in learning

• Essay Editor

Exploring the Impact of Individual Differences on Learning Styles and Academic Achievement

1. introduction.

Though the construct of learning styles has been extensively examined in recent years, its relationship with individual differences and academic achievement is still under-explored. The overwhelming evidence of the negative influence of maladaptive cognitive factors on learning underscores the urgency of examining the phenomena of learning style and maladaptive learning cognitive factors. However, a review of the literature reveals that the bulk of empirical research on learning styles has methodological limitations. This article reviews the literature on individual differences and then examines the hypothesized key individual differences (information style, toleration for ambiguity, and flexibility) that affect an individual's learning style and the impact of one's learning style on scholastic achievement. The findings revealed that information style was the most important aspect of individual differences with a significant effect on students' learning style. Second, students' learning style had an impact on academic achievement, and information style played a key role in affecting academic achievement. It is a well-established fact that individual differences have an influence on learning and academic achievement. The effect of these differences transmits as students' learning style, achievement increases as an outcome of utilizing a preferred learning approach. Thorndike (1913) emphasized the importance of individual difference and took the position that students admitted to higher education could not be fitted into an exact mold. Similarly, Guilford (1959) and Allinson and Hayes (1996) stressed diverse learning abilities; every individual has a unique learning style. Since then, research has been carried out to explore individual traits that lead to differences in learning environments. A growing body of research has been carried out to investigate learning styles from personality in a very wide-ranging sense. However, few studies had addressed the impact of these personality types on learning style. This article aims to investigate the relationship between, first, individual differences and learning style, and second, learning style and academic achievement.

1.1. Background and Significance

The wrong choice and application of learning styles may produce negative effects in students, such as demotivation and difficulty in understanding educational concepts. Nowadays, each university student must know their learning styles to increase their academic performance and satisfaction with the student learning experience in higher education. An alternative teaching system to improve and update students' learning styles is adaptive teaching. This system proposes that teaching may be adapted to the learning styles of the students in a dynamic way through continuous monitoring of the learning process. In an adaptive teaching context, it is necessary to monitor and update the learning styles of students in real time. Under this proposal, updating the learning styles of students is not enough only at the beginning of their studies. Styles and strategies-based education is a promising approach that tries to take advantage of the learning styles and strategies of individuals to increase academic performance and learning retention. The main idea of an education system based on styles and strategies is that students learn well when the instruction is adapted to their learning styles and strategies. The instruction can be based on the specific learning styles of the individuals to leverage what is most comfortable and effective for them, and strategies, or instructional materials and methods that optimize the strengths of students while avoiding the elements that cause problems. School level produces substantial effects on individuals until students arrive at university, so children come with different learning styles, and these are developed by different teachers with different teaching methods. The study of the learning process shows that university students acquire information in different ways, through interaction with the world and through their common attempts to understand it.

1.2. Research Objectives

With the growing awareness for the diversity of today's classrooms, come greater challenges to educators. The aim of this study is to examine two important issues facing instructors: (1) whether appropriate attention is given to variations in learning styles within the student population, which can lead to enhanced learning and comprehension; and (2) whether such attention results in more even levels of academic accomplishment across student classifications. Consequently, this study seeks to answer four main questions: 1. What types of differences exist in learning styles among students? 2. What are the similarities and differences in learning styles for first-year students enrolled in introductory management courses compared to second-year students? 3. What impact do these learning styles have on academic achievement? 4. What can these findings suggest for purposeful design of instructional methods for students representing different learning style preferences? By examining personality traits of students in relation to methods of instruction and looking at the determinants of academic accomplishment, the needs and characteristics of particular groups of students provide a guide for instructional design based on learning style. In such a way, biases are not inherent in the instructional methods chosen.

2. Theoretical Framework

Considering the present study’s major goal, the literature review is structured into two main sections: 1) teaching and learning approaches, and 2) teaching and learning methods. This section focuses on introducing a wide range of concepts that currently mix up pedagogical perspectives and practices worldwide; the methodological components of these concepts are then discussed. Different teaching methods can be seen as representing various areas of each concept. The major goals of this section are: to list widely acknowledged aspects of teaching and learning, to look at correlations between theoretical frames of such aspects, and to find some underlying principles based on the difference and typology of learning and teaching approaches. Individual differences can affect academic performance profoundly. Although it may be a challenge, we can promote students' academic achievements by meeting their various learning styles. The present study aimed to explore associations among the learning styles, learning strategies, and academic performance of medical students. The participants were 720 newly enrolled students in a British-Malaysian medical school. The study used non-matched samples on the assumption that such a decision would result in a more specific comparison. The VARK questionnaire was used to identify a learner's learning style, and a validated questionnaire was employed to measure students' self-reported learning strategies. Academic performance was graded based on how students performed in the subjects. The study results indicated that specific learning style was associated with higher use of specific learning strategies. Learning styles also appeared to influence academic achievement. The findings of the study can help educators adapt teaching plans to students' learning styles and activate their learning progress, especially for those who recently enrolled in challenging courses.

2.1. Individual Differences in Learning Styles

The magnitude of the learning effect of information can change according to factors such as the individual's sex, age, social status, type of employment, educational level, personal productivity, income level of their family, income-generating activities, knowledge in some subjects, and educational status. As a result, all individuals learn information differently. This results from the psychological diversification theory. The diversity of learning and achievement reveals that different learning styles have been revealed and individuals have the opportunity to show the skills and characteristics they apply in different learning styles. In this context, the challenge of overriding diversity is to provide intellectual tasks that allow all students to express their unique learning strengths without labeling them for any loss. Each person is unique, and each individual learns in a different way. This diversity in learning styles has many determinations. Pedagogical, dominant nerve properties, intellectual confrontation related to aesthetics, anxiety, previous experience, intellectual behavior, mental intelligence, mental ability, personality, and sociological variables occur. Although everyone has different learning capabilities, school systems and teaching programs are generally structured in a similar manner and delivered in a one-to-one format for all students.

2.2. Cognitive Theories of Learning

Cognitive theories that recognize the importance of using cognitive learning strategies, an individual's capacity, and level of readiness for learning have emerged as a reaction to behavioral theories. Cognitive learning theory supports the importance of observational learning, meaningful learning, discovery learning, inquiry learning, and problem-solving learning, and focuses on students' individual differences. Ausubel describes cognitive structures such as advance diversity, advance organizer, and operation as: • The relations between new knowledge and existing cognitive structure; • Organize cognitive structures of students in line with the receiving of new knowledge, structuring cognitive structures of students before reflecting the new knowledge. Gagne's model suggests that information should be designed to suit existing cognitive structures. The ability, readiness, interests, and content of the students are the factors determining the existence or the development level of these structures. These theories consist of meaningful concepts, models, visual fractionation models, visual models, and interaction models. More of these models are oriented around the idea that learning is an active process that is under the control of the student, and that the teacher's duty is to facilitate learning. Rogers and Maslow, on the other hand, do not consider students as a passive agent, affirming that they participate actively in all levels of learning and cognitive areas of individual differences; students have the right to say what they want and how they will be educated. In preparing these models, it is assumed that teachers take into consideration individual differences and propose methods that comply with them. They suggest that students are willing, curious, and open to learning if the teaching methods are in line with the student's need for learning and veiling.

3. Methodology

The survey research design was used to describe and explore the communicative language learning styles utilized by the students at the Jordan University of Science and Technology (JUST) in various academic contexts. The survey research makes use of questionnaires to collect data about the features, attitudes, opinions, and possible relationships between them. For that, the variables to be studied were operationalized into questionnaire items which reflected the manifestations of the concepts under study. The population of the study was all the students in English departments in the university. The sample of the study was students from the Colleges of Arts and Sciences. They are taking compulsory courses in applied linguistics, introduction to linguistics, and syntax. The classes usually consist of non-English major students who start learning English at the university level. The number of students in these groups may range from 30 to 50 students. Two questionnaires were designed to measure the four specific dimensions of the developing Learning Style Inventory (i.e. introversion/extroversion, intuition/sensing, thinking/feeling, and judger/perceiver), and the communicative language learning style of the students in language classes. The questionnaires were filled out by the students during class time. Data was collected during two semesters (the 2002/2003 and 2003/2004 academic years). The four dominant dimensions of each of the learning style models developed by Jung, Myers, and Briggs (1943) reflect pairs of counterbalanced preferences. A person may prefer one or the other of each pair.

3.1. Participants

One hundred and thirty-seven middle and high school students participated. 40% were middle school students, and 60% were high school students. Sixty-one percent were male, and 39% were female. The mean age of the participants was 15.56, with a standard deviation of 1.7, ranging from 12 to 18. The participants represented 21 different nationalities and spoke eight different languages. The largest nationality groups were Arabic (28%), Chinese (8%), and Spanish (8%). All students were associated with an international, college-preparatory school located in the Kingdom of Saudi Arabia, where English was the language of instruction. Japanese, Urdu, Korean, Portuguese, and Russian are the other native languages. Six students were not taking Math, and one student was not taking physics. Therefore, their Likert scale scores, including their ratings in the learning styles' section, were not used. The greatest percentage of them (38%) currently attended 11th grade. Seventy-seven percent of the students were attending from 10th to 12th grades, and 23% from 7th to 9th grades. More than half of the students (54%) explained that they would like to study business administration in college.

3.2. Measures

3.2.1. The questionnaire of learning styles LSQ dimensions are based on four subscales which include concrete experience, reflective observation, abstract conceptualization, and active experimentation. Conventional learning is highly practical and experiential. Both of those and reflective observation are considered to be the two sides of perception, while abstract conceptualization and active experimentation are considered the two forms of perception. Concrete experience and active experimentation capabilities are vice versa of reflective observation and abstract conceptualization. It means that people prefer one of the two sides of perception in the learning process. 3.2.2. Kolb cycle inventory The Kolb cycle inventory defines learning styles through the four basic stages of learning: doing and feeling, watching and thinking, looking back and learning, change, and planning. These four learning styles are: sensibility, reflection beyond the consideration, abstract conceptualization, and active examination; Kolb assumes that sensation and reflection are the two aspects of perception and that abstract conceptualization and active experimentation are the two forms of processing. When a person is taken from the perception and from the treatment aspect, these four styles of learning refer to two different learning processes. 3.2.3. Opinions variable The purpose of this study was to investigate the individual aspects that affect the learning style and academic achievement. Some personal characteristics, such as gender, age, and senior figures, were personal-controlled variables, and various research projects had different effects on the scores of formal studies in psychological field studies. These moderating variables are particularly important when the program aims to affect the predetermined outcome variable.

To examine the impact of individual differences on both student learning styles and academic performance, the following analysis tools were used: frequencies and percentages, with Chi-square tests to test for differences among groups of equal variances of student learning styles classified with the help of Cluster analysis output, and among groups of student achievement (35%) using a convenient pseudo-set of centiles independently and taking into account the specific proportion of students across seven faculties. This approach also allowed us to overcome several limitations. First of all, it enabled the direct interpretation of the influence of both "background" and "environmental" factors with significant effects on students' levels of performance. Moreover, the incorporation into analysis over any pseudo-set of achievements, independently calculated as "% of students achieving results by various levels", helped to evade the "gender biased" phenomena like female "half marks" and male "round figure" marks. Table 1 documented data of (EJM and IQM) on student results offered a cautious statement that in an environment of high competition, inconsistencies in average student academic achievement are present. For California Polytechnic College, as one might anticipate, a very narrow cluster of active students tend to gravitate about 80% of the students scoring with average, good, or excellent marks.

4.1. Correlation Analysis

The results reported in Table 1 depicted a significant relationship between the 20 dimensions of this study. The significant positive correlations were established among the dimensions of verbal-linguistic intelligence, logical-mathematical intelligence, intrapersonal intelligence, and interpersonal skills. This implies that there were close interrelations among these individual differences. This result implies that students with one type of the characteristics of these individual differences are likely to demonstrate other characteristics simultaneously. For example, it is predicted that the students who reveal verbal-linguistic intelligence are likely to reveal logical-mathematical intelligence or intrapersonal intelligence or interpersonal skills, which is likely to point out the existence of the general factor for individual differences. Consequently, the prediction is that the correlation coefficients between pairs of these 20 dimensions in the present investigation would be positive significant. The examination of Table 1 reveals the correlation values and significances. The results have indicated that the reckoned correlation coefficients across the dimensions of physical values, environment sensitivity, religious values and inquiring mind, logical-mathematical intelligence, intrapersonal intelligence, and interpersonal skills have been demonstrated, as hypothesized, significant positive interrelations with all of the abilities or skills. Many of the significant coefficient of correlations chilled to decline. This result suggests that students who reveal other dimensions would not express abilities or skills. Remarkably, the abilities or skills did not tend to solidarity against the dimensions of social values and type of schools, these dimensions were not significantly linked with the scores of the abilities and skills. In this respect, all the other dimensions typically co-occur to existent variables under consideration. The assumption has been confirmed as the students do not appear to exclusively demonstrate the alliance with particular abilities or skills. This means that they have risen to magnify variations in the scores of other abilities or skills.

5. Discussion

Thus, the research explored the impact of learning styles and academic achievement, as well as the effects of various individual difference variables on learning style and achievement. In general, it emphasized the importance of considering a more complex model of education, which takes into account a wider range of cognitive, emotional, social, and motivational development, and the relationship of education seemingly with biographical diversity. Although the data is not contemporary, students in the age group covered are potentially more at risk of underachievement and drop out, and might be the subject of developmental teaching and more active curriculum and support. Institutions need to take into account the changing composition and needs of their students. Students varied in a wide range of factors, some critical to their learning experiences, such as motivation, emotional responses to learning, and the teaching environment. The results show that a single style does not cater to such a wide diversity of students, and classes should not expect to teach in a single way. They should also adopt more interactive models of curriculum and learning, based on team tuition, flexible and open access methods of information propagation, use of VLEs to reinforce and extend learning, and so on. Teachers were the critical change factor, and this had to do more with better curriculum design, innovative development of study materials, and promotion of active learning strategies. The authors took a pragmatic approach. Their conclusion was that equal status for different teaching strategies would give partnership in providing diverse and responsive learning environments. Adaptive tuition methods enabled those students to exceed expectations while nurturing a high-quality, stimulating learning experience for all.

5.1. Implications for Education

Recent approaches to the teaching process are steered by the principles of constructive and collaborative learning strategies and their direct implementation on educational practices. However, empirical evidence shows that applying such strategies frequently results in poor educational outcomes such as failure and dropout rates, and most importantly, in a high inter-individual variation in students' achievement. This variability is reflected in the fact that not all learners show proportional performance improvements with the current instruction methods, abandoning the ultimate goal of personalized education, able to match diverse learners' profiles. Although the underlying mechanisms remain only partly understood, it may be assumed that part of the poor educational outcomes achieved in real life results from the low level of understanding that exists on the individual differences shaping students' profiles and the possible consequences they may have in engaging the learning process. In this context, the identification of the factors and conditions that affect how students might best learn is a fundamental step in the development of educational strategies and curricula for the enhancement of students' performance. However, such indicators are usually not considered and are indeed important for identifying the most effective pedagogic strategies to be adopted. In addition to informing us about priors of learning outcomes, the identification of educational indicators also has a practical impact for educators in terms of behavioral drivers.

Related articles

The impact of qualitative and quantitative research methods on curriculum vitae evaluations.

1. Introduction The question of how to evaluate a helper - be it an applicant for a job, an employee who should be promoted or a response team member who should be sent to take care of customers - is an important question in many areas of life. Methods to evaluate the qualities of a person are used in academic tenure decisions, in secondary school and college admissions, in job applications, and in many more contexts. Such decisions are important for a principal who wants to select the best n o ...

The Impact of Academic Stress on Student Health and Well-being

1. Introduction College can be a stressful experience. Not only do students need to learn academic material, they must also find ways to more effectively manage their day-to-day living and study habits, and develop skills relevant to keeping themselves healthy and free of tuberculosis. While stress is generally a part of college life, with students encountering stressors, for example, through coursework or examinations, some common sources of stress are discussed below. A notable increase over ...

The Impact of Social Media on Adolescent Behavior During Critical Developmental Stages

1. Introduction Adolescent development is motivated and influenced by cognitive processes within numerous areas that include: the brain, self-ranking, preferred risk-taking activities, and promotion of novelty. These are areas that are usually influenced by serotonergic activity, which is impacted more in adolescence. Serotonin is helpful for brain regional connectivity, which enables coordinated brain development. Without this development, overall integration of frontal cortex and subcortical ...

The Significance of Citing Sources in Academic Writing and the Ramifications of Plagiarism

1. Introduction Citations to sources are an essential element in scholarly writing. In purporting to inform the audience on a scholarly point, academic writers claim potency and convincingness for their assertions by referencing sources that confirm their point of view. A distinction is made between findings, which are based on ancient evidence and should be cited appropriately, and novel ideas or theories, which are the product of the author and an absence of evidence exists to support the vie ...

The Impact of Economic Policies on the Middle Class: Challenges and Opportunities

1. Introduction During the thirty years after World War II, the American middle class grew, multiplied, and became more prosperous. But in the half-century since then, progress for the middle class has slowed and, for many, even stopped. And now, in the current economic crisis, families in the middle have been affected more severely than most groups and are finding it difficult to bounce back. The experience of the GI Bill has left a historical example that shows how effective policy can be for ...

Exploring the Impact of Duración (Duration) on Language Acquisition in Bilingual Education

1. Introduction Educational systems around the world have adopted bilingual education in pursuit of the linguistic empowerment of their students. Bilingual education aims to provide proficiency in two or more languages, with specific objectives that have been described as follows: "For a significant period of time, it is the primary medium of instruction; students should achieve high levels of proficiency in both their native tongue and a second language". As some have recognized, "bilingualism ...

The Importance of Improving Writing Skills in Academic and Professional Success

1. Introduction I thoroughly believe that developing students' writing skills translates into driving student engagement in the classroom. This is reflected not only in students' writing skills but also in their ability to communicate and express their ideas. In my experience as a teacher, regardless of the context where I have taught, I have noticed that students' failures in the programs' academic or professional activities and projects are partly due to weak writing skills. This happens even ...

Assessing the Impact of Technology on Modern Education

1. Introduction to Technology in Education The rapid pace of technological growth has had a major impact on the educational system of the United States. Many would argue that the effects of technology on education are many. Proponents argue that computers, the internet, and technology all promote improved achievement scores and productivity. Opponents argue that the internet serves as a means of enabling cheating and increases racial, social, and geographical inequities in education. The purpos ...

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Review Article
• Open access
• Published: 12 January 2017

Individual differences in the learning potential of human beings

• Elsbeth Stern 1  

npj Science of Learning volume  2 , Article number:  2 ( 2017 ) Cite this article

65k Accesses

22 Citations

67 Altmetric

Metrics details

• Human behaviour

To the best of our knowledge, the genetic foundations that guide human brain development have not changed fundamentally during the past 50,000 years. However, because of their cognitive potential, humans have changed the world tremendously in the past centuries. They have invented technical devices, institutions that regulate cooperation and competition, and symbol systems, such as script and mathematics, that serve as reasoning tools. The exceptional learning ability of humans allows newborns to adapt to the world they are born into; however, there are tremendous individual differences in learning ability among humans that become obvious in school at the latest. Cognitive psychology has developed models of memory and information processing that attempt to explain how humans learn (general perspective), while the variation among individuals (differential perspective) has been the focus of psychometric intelligence research. Although both lines of research have been proceeding independently, they increasingly converge, as both investigate the concepts of working memory and knowledge construction. This review begins with presenting state-of-the-art research on human information processing and its potential in academic learning. Then, a brief overview of the history of psychometric intelligence research is combined with presenting recent work on the role of intelligence in modern societies and on the nature-nurture debate. Finally, promising approaches to integrating the general and differential perspective will be discussed in the conclusion of this review.

Similar content being viewed by others

Uniquely human intelligence arose from expanded information capacity

Genetic variation, brain, and intelligence differences

Change by challenge: A common genetic basis behind childhood cognitive development and cognitive training

Human learning and information processing.

In psychology textbooks, learning is commonly understood as the long-term change in mental representations and behavior as a result of experience. 1 As shown by the four criteria, learning is more than just a temporary use of information or a singular adaption to a particular situation. Rather, learning is associated with changes in mental representations that can manifest themselves in behavioral changes. Mental and behavioral changes that result from learning must be differentiated from changes that originate from internal processes, such as maturation or illness. Learning rather occurs as an interaction with the environment and is initiated to adapt personal needs to the external world.

From an evolutionary perspective, 2 living beings are born into a world in which they are continuously expected to accomplish tasks (e.g., getting food, avoiding threats, mating) to survive as individuals and as species. The brains of all types of living beings are equipped with instincts that facilitate coping with the demands of the environment to which their species has been adapted. However, because environments are variable, brains have to be flexible enough to optimize their adaptation by building new associations between various stimuli or between stimuli and responses. In the case of classical conditioning, one stimulus signals the occurrence of another stimulus and thereby allows for the anticipation of a positive or negative consequence. In the case of operant conditioning, behavior is modified by its consequence. Human beings constantly react and adapt to their environment by learning through conditioning, frequently unconsciously. 1

However, there is more to human learning than conditioning, which to the best of our knowledge, makes us different from other species. All living beings must learn how to obtain access to food in their environment, but only human beings cook and have invented numerous ways to store and conserve their food. While many animals run faster than humans and are better climbers, the construction and use of vehicles or ladders is unique to humans. There is occasional evidence of tool use among non-human species passed on to the next generation, 3 , 4 but this does not compare to the tools humans have developed that have helped them to change the world. The transition from using stonewedges for hunting to inventing wheels, cars, and iPhones within a time period of a few thousand years is a testament to the unique mental flexibility of human beings given that, to the best of our knowledge, the genes that guide human brain development have not undergone remarkable changes during the last 50,000 years. 5 This means that as a species, humans are genetically adapted to accomplish requirements of the world as it existed at approximately 48,000 BC. What is so special about human information processing? Answers to this question are usually related to the unique resource of consciousness and symbolic reasoning abilities that are, first and foremost, practiced in language. Working from here, a remarkable number of insights on human cognition have been compiled in the past decades, which now allow for a more comprehensive view of human learning.

Human learning from a general cognitive perspective

Learning manifests itself in knowledge representations processed in memory. The encoding, storage, and retrieval of information have been modeled in the multi-store model of human memory depicted in Fig.  1 . 6 Sensory memory is the earliest stage of processing the large amount of continuously incoming information from sight, hearing, and other senses. To allow goal-directed behavior and selective attention, only a fractional amount of this information passes into the working memory, which is responsible for temporarily maintaining and manipulating information during cognitive activity. 7 , 8 Working memory allows for the control of attention and thereby enables goal-directed and conscious information processing. It is the gatekeeper to long-term memory, which is assumed to have an unlimited capacity. Here, information acquired through experience and learning can be stored in different modalities as well as in symbol systems (e.g., language, script, mathematical notation systems, pictorials, music prints).

A model of human information processing, developed together with Dr. Lennart Schalk

The multi-store model of human information processing is not a one-way street, and long-term memory is not to be considered a storage room or a hard-disk where information remains unaltered once it has been deposited. A more appropriate model of long-term memory is a self-organizing network, in which verbal terms, images, or procedures are represented as interlinked nodes with varying associative strength. 9 Working memory regulates the interaction between incoming information from sensory memory and knowledge activated from long-term memory. Very strong incoming stimuli (e.g., a loud noise or a harsh light), which may signal danger, can interrupt working memory activities. For the most part, however, working memory filters out irrelevant and distracting information to ensure that the necessary goals will be achieved undisturbed. This means that working memory is continuously selecting incoming information, aligning it with knowledge retrieved from long-term memory, and preparing responses to accomplishing requirements demanded by the environment or self-set goals. Inappropriate and unsuitable information intruding from sensory as well as from long-term memory has to be inhibited, while appropriate and suitable information from both sources has to be updated. 8 The strength with which a person pursues a particular goal has an impact on the degree of inhibitory control. In case of intentional learning, working memory guards more against irrelevant information than in the case of mind wandering. Less inhibitory control makes unplanned and unintended learning possible (i.e., incidental learning).

These working memory activities are permanently changing the knowledge represented in long-term memory by adding new nodes and by altering the associative strength between them. The different formats knowledge can be represented in are listed in Fig.  1 ; some of them are more closely related to sensory input and others to abstract symbolic representations. In cognitive psychology, learning is associated with modifications of knowledge representations that allow for better use of available working memory resources. Procedural knowledge (knowing how) enables actions and is based on a production-rule system. As a consequence of repeated practice, the associations between these production rules are strengthened and will eventually result in a coordinated series of actions that can activate each other automatically with a minimum or no amount of working memory resources. This learning process not only allows for carrying out the tasks that the procedural knowledge is tailored to perform more efficiently, but also frees working memory resources that can be used for processing additional information in parallel. 10 , 11 , 12

Meaningful learning requires the construction of declarative knowledge (knowing that), which is represented in symbol systems (language, script, mathematical, or visual-spatial representations). Learning leads to the regrouping of declarative knowledge, for instance by chunking multiple unrelated pieces of knowledge into a few meaningful units. Reproducing the orally presented number series “91119893101990” is beyond working memory capacity, unless one detects two important dates of German history: the day of the fall of the Berlin Wall: 9 November 1989 and the day of reunification: 3 October 1990. Individuals who have stored both dates and can retrieve them from long-term memory are able to chunk 14 single units into two units, thereby freeing working memory resources. Memory artists, who can reproduce dozens of orally presented numbers have built a very complex knowledge base that allows for the chunking of incoming information. 13

Learning also manifests itself in the extension of declarative knowledge using concept formation and inferential reasoning. Connecting the three concepts of “animal, produce, milk” forms a basic concept of cow. Often, concepts are hierarchically related with superordinate (e.g., animal) and subordinate (e.g., cow, wombat) ordering. This provides the basis for creating meaningful knowledge by deductive reasoning. If the only thing a person knows about a wombat is that it is an animal, she can nonetheless infer that it needs food and oxygen. Depending on individual learning histories, conceptual representations can contain great variations. A farmer’s or a veterinarian’s concept of a cow is connected to many more concepts than “animal, produce, milk” and is integrated into a broader network of animals. In most farmers’ long-term memory, “cow” might be strongly connected to “pig”, while veterinarians should have particularly strong links to other ruminants. A person’s conceptual network decisively determines the selection and representation of incoming information, and it determines the profile of expertise. For many academic fields, first and foremost in the STEM area (Science, Technology, Engineering, Mathematics), it has been demonstrated that experts and novices who use the same words may have entirely different representations of their meaning. This has been convincingly demonstrated for physics and particularly in the area of mechanics. 14 Children can be considered universal novices; 15 therefore, their everyday concepts are predominantly based on characteristic features while educated adults usually consider defining features, 16 , 17 , 18 as the example of “island” demonstrates. For younger children, it primarily refers to a warm place where one can spend ones’ holidays. In contrast, adults’ concept of island does refer to a tract of land that is completely surrounded by water but not large enough to be considered a continent.

The shift from characteristic to defining features is termed “conceptual change”, 16 and promoting this kind of learning is a major challenge for school education. Students’ understanding of central concepts in an academic subject can undergo fundamental changes (e.g., the concept of weight in physics). Younger elementary school children often agree that a pile of rice has weight, but they may also deny that an individual grain of rice has weight at all. This apparently implausible answer is understandable given that younger children consider the concepts of “weight” and “being heavy” as equivalent. As such, children tend to agree that a grain of rice has weight if it is put on an ant’s back. 16 As a consequence of their education, students usually understand that an object’s weight is determined with the assistance of scales and not necessarily by personal sensation. However, representing weight as the property of an object is still not compatible with scientific physics in the Newtonian sense by which weight is conceptualized as a relation between objects. Understanding weight in this sense requires an interrelated network of knowledge, including the concepts of force, gravity, and mass (among others).

As a result of classroom instruction, students are expected to acquire procedural and conceptual knowledge of the subjects they were taught. While procedures emerge as a function of repetition and practice, the acquisition of advanced concepts, which are consistent with state of the art science, is less straightforward. 14 , 19 To support this kind of conceptual learning, insights from cognitive learning research have been integrated into educational research and are increasingly informing classroom practice. Several instructional methods have been developed and evaluated that support students in restructuring and refining their knowledge and thereby promote appropriate conceptual understanding, including self-explanations, 20 contrasting cases, 21 , 22 and metacognitive questions. 23 Cognitive research has also informed the development of the “taxonomy of learning objects”. 24 This instrument is widely employed for curriculum development and in teacher training programs to support the alignment of content-specific learning goals, means of classroom practice, and assessment. The taxonomy acknowledges the distinction between procedural and conceptual knowledge and includes six cognitive processes (listed in Fig.  1 ) that describe how knowledge can be transformed into observable achievement.

How core knowledge innate to humans can meet with academic learning

What makes humans efficient learners, however, goes beyond general memory functions discussed so far. Similar to other living beings, humans do not enter the world as empty slates 2 but are equipped with so-called core knowledge (Fig.  1 ). Evidence for core knowledge comes from preferential looking experiments with infants who are first habituated to a particular stimulus or scenario. Then, the infant is shown a second scenario that differs from the first in a specific manner. If the time he or she looks at this stimulus exceeds the looking-time at the end of the habituation phase of the first stimulus, this suggests that the infant can discriminate between the stimuli. This paradigm helps to determine whether infants detect violations of principles that underlie the physical world, such as the solidity of objects, where an object cannot occupy the same space as another object. 25 , 26 Core knowledge, which allows privileged learning and behavioral functioning with little effort, also guides the unique human ability of symbolic communication and reasoning, first and foremost, langue learning. 27 , 28 It is uncontested that humans are born with capacities for language learning, which includes the awareness of phonological, grammatical, and social aspects of language. 4 , 29 , 30

Core knowledge can serve as a starting point for the acquisition of content knowledge that has emerged as a result of cultural development. This has been examined in detail for numerical and mathematical reasoning. Two core systems have been detected in infants. As early as at 6 months of age, infants show an ability for the approximate representations of numerical magnitude, which allow them to discriminate two magnitudes depending on their ratio. 31 At the same age, the system of precise representations of distinct individuals allows infants to keep track of changes in small sets of up to three elements. 32 Mathematical competencies emerge as a result of combining both core systems and linking them to number words provided by the respective culture. 33 The Arabic place value number system, which is now common in most parts of the world, was only developed a few 100 years ago. Only after the number “0” had made its way from India via the Arabic countries to Europe were the preconditions for developing our decimal system available. 34 The Arabic number system opened up the pathway to academic mathematics. Cultural transformations based on invented symbol systems were the key to advanced mathematics. Today’s children are expected to understand concepts within a few years of schooling that took mankind centennials to develop. Central content areas in mathematics curricula of high schools, such as calculus, were only developed less than three centuries ago. 35 Given the differences between the Arabic and the Roman number systems, children born 2000 years ago could not make use of their numerical core knowledge in the same way today’s children can.

Core knowledge about navigation is meant to guide the acquisition of geometry, an area involved in numerous academic fields. 36 , 37 The cornerstone of cultural development was the invention of writing, in which language is expressed by letters or other marks. Script is a rather recent cultural invention, going back approximately 5,000 years, whereas the human genome emerged approximately 50,000 years ago. 38 Clearly, unlike oral language, humans are not directly prepared for writing and reading. Nonetheless, today, most 6-year-old children become literate during their 1st years of schooling without experiencing major obstacles. Human beings are endowed with the many skills that contribute to the ability to write and read, such as, first and foremost, language as well as auditory and visual perception and drawing. These initially independent working resources were coopted when script was invented, and teaching children to write and read at school predominantly means supporting the development of associations among these resources. 39

Part of the core knowledge innate to humans has also been found in animals, for instance numerical knowledge and geometry, but to the best of our knowledge, no other animals have invented mathematics. 40 Only humans have been able to use core knowledge for developing higher order cognition, which serves as a precondition for culture, technology, and civilization. Additionally, the unique function of human working memory is the precondition for the integration of initially independent representational systems. However, the full potential of working memory is not in place at birth, but rather matures during childhood and undergoes changes until puberty. 41 Children under the age of two are unable to switch goals 42 and memorize symbol representations appropriately. 43

To summarize what has been discussed so far, there are two sources for the exceptional learning capacity of humans. The first is the function of working memory as a general-purpose resource that allows for holding several mental representations simultaneously for further manipulation. The second is the ancient corpus of the modularized core knowledge of space, quantities, and the physical and social world. Working memory allows for the connection of this knowledge to language, numerals, and other symbol systems, which provides the basis for reasoning and the acquisition of knowledge in academic domains, if appropriate learning opportunities are provided. Both resources are innate to human beings, but they are also sources of individual differences, as will be discussed in the following sections.

Learning potentials are not alike among humans: the differential perspective

In the early twentieth century, a pragmatic need for predicting the learning potential of individuals initiated the development of standardized tests. The Frenchman Alfred Binet, who held a degree in law, constructed problems designed to determine whether children who did not meet certain school requirements suffered from mental retardation or from behavioral disturbances. 44 He asked questions that still resemble items in today’s intelligence tests; children had to repeat simple sentences and series of digits forwards and backwards as well as define words such as “house” or “money”. They were asked in what respect a fly, an ant, a butterfly and a flea are alike, and they had to reproduce drawings from memory. William Stern, an early professor of psychology at the newly founded University of Hamburg/Germany, intended to quantify individual differences in intelligence during childhood and adolescence by developing the first formula for the intelligence quotient (IQ): 45 IQ = Mental age/chronological age*100. Mental age refers to the average test score for a particular age group; this means that a 6-year-old child would have an IQ = 133 if their test score was equivalent to the mean score achieved in the group of 8-year-olds. From adolescence on, however, the average mental age scores increasingly converge, and because of the linear increase in chronological age, the IQ would decline—a trend that obviously does not match reality.

Psychologists from the United States, specifically headed by the Harvard and later Yale professor Robert Yerkes, decided to look at a person’s score relative to other people of the same age group. The average test score was assigned to an IQ = 100 by convention, and an individual’s actual score is compared to this value in terms of a standard deviation, an approach that has been retained to this day. World War I pushed the development of non-verbal intelligence tests, which were used to select young male immigrants with poor English language skills for military service. 46 In the UK, the educational psychologist Cyril Burt promoted the use of intelligence tests for assigning students to the higher academic school tracks. 47 Charles Spearman from the University College London was among the first to focus on the correlations between test items based on verbal, numerical, or visual-spatial content. 48 The substantial correlations he found provided evidence for a general intelligence model (factor-g), which has been confirmed in the following decades by numerous studies performed throughout the world. 49

The high psychometric quality of the intelligence tests constructed in different parts of the world by scientists in the early decades of the twentieth century have influenced research ever since. In 1923, Edward Boring, a leading experimental psychologist concluded, “Intelligence is what the tests test. This is a narrow definition, but it is the only point of departure for a rigorous discussion of the tests. It would be better if the psychologists could have used some other and more technical term, since the ordinary connotation of intelligence is much broader. The damage is done, however, and no harm need result if we but remember that measurable intelligence is simply what the tests of intelligence test, until further scientific observation allows us to extend the definition.”(ref. 50 , p. 37). More than 70 years later, psychologists widely agreed on a definition for intelligence originally offered by Linda Gottfredsonin 1997: “Intelligence is a very general mental capability that, among other things, involves the ability to reason, plan, solve problems, think abstractly, comprehend complex ideas, learn quickly, and learn from experience. It is not merely book learning, a narrow academic skill, or test-taking smarts. Rather, it reflects a broader and deeper capability for comprehending our surroundings—‘catching on,’ ‘making sense’ of things, or ‘figuring out’ what to do” (ref. 51 , p. 13). This definition is in line with the substantial correlations between intelligence test scores and academic success, 52 whereas correlations with measures of outside-school success, such as income or professional status, are lower but still significant. 53 , 54 Numerous longitudinal studies have revealed that IQ is a fairly stable measure across the lifespan, which has been most convincingly demonstrated in the Lothian Birth Cohorts run in Scotland. Two groups of people born in 1921 and 1936 took a test of mental ability at school when they were 11 years old. The correlation with IQ tests taken more than 60 years later was highly significant and approached r  = .70 (ref. 55 ). The same data set also demonstrated a substantial long-term impact of intelligence on various factors of life success, among them career aspects, health, and longevity. 56

Intelligence tests scores have proven to be objective, reliable, and valid measures for predicting learning outcome and more general life success. At the same time, the numerous data sets on intelligence tests that were created all over the world also contributed to a better understanding of the underlying structure of cognitive abilities. Although a factor g could be extracted in almost all data sets, correlations between subtests varied considerably, suggesting individual differences beyond general cognitive capabilities. Modality factors (verbal, numerical, or visual spatial) have been observed, showing increased correlations between tests based on the same modality, but requiring different mental operations. On the other hand, increased correlations were also observed between tests based on different modalities, but similar mental operations (e.g., either memorizing or reasoning). The hierarchical structure of intelligence, with factor g on the top and specific factors beneath, was quite obvious from the very beginning of running statistical analyses with intelligence items. Nonetheless, it appeared a major challenge for intelligence researchers to agree on a taxonomy of abilities on the second and subsequent levels. In 1993, John Carroll published his synthesis of hundreds of published data sets on the structure of intelligence after decades of research. 57 In his suggested three-stratum model, factor g is the top layer, with the middle layer encompassing broader abilities such as comprehension knowledge, reasoning, quantitative knowledge, reading and writing, and visual and auditory processing. Eighty narrower abilities, such as spatial scanning, oral production fluency, and sound discrimination, are located in the bottom layer. To date, Carroll’s work is considered the most comprehensive view of the structure of individual variations in cognitive abilities. 58 However, the interpretation of factor g is still under discussion among scientists. Factor g could be a comprehensive characteristic of the brain that makes information processing generally more or less efficient (top-down-approach). Existing data sets, however, are also compatible with a model of intelligence according to which the human brain is comprised of a large number of single abilities that have to be sampled for mental work (bottom-up approach). In this case, factor g can be considered a statistical correlate that is an emerging synergy of narrow abilities. 59

Genetic sources of individual differences in intelligence

From studies with identical and fraternal twins, we know that genetic differences can explain a considerable amount of variance in IQ. The correlation between test scores of identical twins raised together approaches r  = .80 and thereby is almost equal to the reliability coefficient of the respective test. On the other hand, IQ-correlations between raised-together same-sex fraternal twins are rarely higher than .50, a value also found for regular siblings. Given that the shared environment for regular siblings is lower than for fraternal twins, this result qualifies the impact of environmental factors on intelligence. The amount of genetic variance is judged in statistical analyses based on the difference between the intra-pair correlations for identical and fraternal twins. 60 High rates of heritability, however, do not mean that we can gauge a person’s cognitive capabilities from his or her DNA. The search for the genes responsible for the expression of cognitive capabilities has not yet had much success, despite the money and effort invested in human genome projects. It is entirely plausible that intelligence is formed by a very large number of genes, each with a small effect, spread out across the entire genome. Moreover, these genes seem to interact in very complicated ways with each other as well as with environmental cues. 61

An entirely false but nonetheless still widespread misunderstanding is to equate “genetic sources” with “inevitability” because people fail to recognize the existence of reaction norms, a concept invented in 1909 by the German biologist, Richard Woltereck. Reaction norms depict the range of phenotypes a genotype can produce depending on the environment. 62 For some few physiological individual characteristics (e.g., the color of eyes) the reaction norm is quite narrow, which means gene expression will rarely be affected by varying environments. Other physiological characteristics, such as height, have a high degree of heritability and a large reaction norm. Whether an individual reaches the height made possible by the genome depends on the nutrition during childhood and adolescence. In a wealthy country with uniform access to food, average height will be larger than in a poor country with many malnourished inhabitants. However, within both countries, people vary in height. The heritability in the wealthy country can be expected to approach 100% because everybody enjoyed sufficient nutrition. In contrast, in the poor country, some were sufficiently nourished and, therefore, reached the height expressed by their genome, while others were malnourished and, therefore, remained smaller than their genes would have allowed under more favorable conditions. For height, the reaction norm is quite large because gene expression depends on nutrition during childhood and adolescence. This explains the well-documented tendency for people who have grown up in developed countries to become progressively taller in the past decades.

The environment regulates gene expression, which means that instead of “nature vs. nurture”, a more accurate phrase is “nature via nurture”. 63 The complex interaction between genes and environment can also explain the fact that heritability of intelligence increases during the lifespan. 61 This well-established finding is a result of societies in which a broad variety of cognitive activities available in professional and private life enable adults more than children to actively select special environments that fit their genes. People who have found their niche can perfect their competencies by deliberate learning.

In the first decades of developing intelligence tests, researchers were naive to the validity of non-verbal intelligence; so-called culture-free or culture-fair tests, based on visual-spatial material such as mirror images, mazes or series and matrices of geometric figures, were supposed to be suitable for studying people of different social and cultural levels. 64 This is now considered incorrect because in the meantime, there has been overwhelming evidence for the impact of schooling on the development of intelligence and the establishment and stabilization of individual differences. Approximately 10 years of institutionalized education is necessary for the intelligence of individuals to approach its maximum potential. 65 , 66 , 67

Altogether, twin and adoption studies suggest that 50–80% of IQ variation is due to genetic differences. 61 This relatively large range in the percentage across different studies is due to the heritability of intelligence in the population studied, specifically, the large reaction norm of the genes giving rise to the development of intelligence. Generally, the amount of variance in intelligence test scores explained by genes is higher the more society members have access to school education, health care, and sufficient nutrition. There is strong evidence for a decrease in the heritability of intelligence for children from families with lower socioeconomic status (SES). For example, lower SES fraternal twins resembled each other more than higher SES ones, indicating a stronger impact of shared environment under the former condition. 68 In other words, because of the less stimulating environment in lower SES families, the expression of genes involved in the development of intelligence is likely to be hampered. Although it may be counterintuitive at first, this suggests that a high heritability rate of intelligence in a society is an indicator of economic and educational equity. Additionally, this means that countries that ensure access to nutrition, health care, and high quality education independent of social background enable their members to develop their intelligence according to their genetic potential. This was confirmed by a meta-analysis on interactions between SES and heritability rate. While studies run in the United States showed a positive correlation between SES and heritability rate, studies from Western Europe countries and Australia with a higher degree of economic and social equality did not. 69 , 70

Cognitive processes behind intelligence test scores: how individuals differ in information processing

In the first part of this paper, cognitive processes were discussed that, in principle, enable human beings to develop the academic competencies that are particularly advantageous in our world today. In the second part, intelligence test scores were shown to be valid indicators of academic and professional success, and differences in IQ were shown to have sound genetic sources. Over many decades, research on cognitive processes and psychometric intelligence has been developing largely independently of one another, but in the meantime, they have converged. Tests that were developed to provide evidence for the different components of human cognition revealed large individual differences and were substantially correlated with intelligence tests. Tests of memory function were correlated with tests of factor g. Sensory memory tests have shown that the exposure duration required for reliably identifying a simple stimulus (inspection time) is negatively correlatedwith intelligence. 71 For working memory, there is a large body of research indicating substantial relationships between all types of working memory functions and IQ, with average correlations >.50 (refs 72 , 73 , 74 ). In these studies, working memory functions are measured by speed tasks that require goal-oriented active monitoring of incoming information or reactions under interfering and distracting conditions. Neural efficiency has been identified as a major neural characteristic of intelligence; more intelligent individuals show less brain activation (measured by electroencephalogram or functional magnetic resonance imaging) when completing intelligence test items 75 , 76 as well as working memory items. 77 Differences in information-processing efficiency were already found in 4-month-old children. Most importantly, they could predict psychometric intelligence in 8-year-old children. 78

These results clearly suggest that a portion of individual differences can be traced back to differences in domain-general cognitive competencies. However, psychometric research also shows that individual differences do exist beyond factor g on a more specific level. Differences in numerical, language, and spatial abilities are well established. Longitudinal studies starting in infancy suggest that sources of these differences may be traced back to variations in core knowledge. Non-symbolic numerical competencies in infancy have an impact on mathematical achievement. 79 Similar long-term effects were found for other areas of core knowledge, 80 particularly language. 81

Endowed with general and specific cognitive resources, human beings growing up in modern societies are exposed to informal and formal learning environments that foster the acquisition of procedural as well as declarative knowledge in areas that are part of the school curriculum. Being endowed with genes that support efficient working memory functions and that provide the basis for usable core knowledge allows for the exploitation of learning opportunities provided by the environment. This facilitates the acquisition of knowledge that is broad as well as deep enough to be prepared for mastering the, as of yet, unknown demands of the future. 18 Regression analyses based on longitudinal studies have revealed that the confounded variance of prior knowledge and intelligence predicts learning outcome and expertise better than each single variable. 82 , 83 , 84 Importantly, no matter how intelligent a person is, gaining expertise in a complex and sophisticated field requires deliberate practice and an immense investment of time. 85 However, intelligence differences will come into play in the amount of time that has to be invested to reach a certain degree of expertise. 86 Moreover, intelligence builds a barrier to content areas in which a person can excel. As discussed in the first part of this paper, some content areas—first and foremost from STEM fields—are characterized by abstract concepts mainly based on defining features, which are themselves integrated into a broader network of other abstract concepts and procedures. Only individuals who clearly score above average on intelligence tests can excel in these areas. 84 , 87 For individuals who were fortunate enough to attend schools that offered high-quality education, intelligence and measures of deep and broad knowledge are highly correlated. 88 , 89 A strong impact of general intelligence has also been shown for university entrance tests such as the SAT, which mainly ask for the application of knowledge in new fields. 90 , 91 Societies that provide uniform access to cognitively stimulating environments help individuals to achieve their potential but also bring to bear differences in intelligence. Education is not the great equalizer, but rather generates individual differences rooted in genes.

Omrod, J. E. Human Learning (Pearson, 2012).

Cosmides, L. & Tooby, J. Evolutionary psychology: New perspectives on cognition and motivation. Annu. Rev. Psychol. 64 , 201–229 (2013).

Article   PubMed   Google Scholar  

Spelke, E. S. in Language in Mind: Advances in the Investigation of Language and Thought (eds Gentner, D. & Goldin-Meadow, S.) (MIT Press, 2003).

Tomasello, M. A Natural History of Human Thinking (Harvard University Press, 2014).

Pääbo, S. The diverse origins of the human gene pool. Nat. Rev. Genet. 16 , 313–314 (2015).

Atkinson, R. & Shiffrin, R. in The Psychology of Learning and Motivation: Advances in Research and Theory (eds Spence, K. & Spence, J.) Vol. 2 (Academic Press, 1968).

Baddeley, A. Working memory: looking back and looking forward. Nat. Rev. Neurosci. 4 , 829–839 (2003).

Article   CAS   PubMed   Google Scholar  

Barrouillet, P., Portrat, S. & Camos, V. On the law relating processing to storage in working memory. Psychol. Rev. 118 , 175–192 (2011).

Kintsch, W. Comprehension: A Paradigm for Cognition (Cambridge University Press, 1998).

Anderson, J. R. et al. An integrated theory of the mind. Psychol. Rev. 111 (4), 1036–1060 (2004).

Goldwater, M., Schalk, L. Relational categories as a bridge between cognitive and educational research. Psychol. Bull. 729–757 (2016).

Schalk, L., Saalbach, H. & Stern, E. Approaches to foster transfer of formal principles: which route to take? PLoS ONE 11 (2), e0148787, doi: 10.1371/journal.pone.0148787 (2016).

Article   PubMed   PubMed Central   Google Scholar  

Chase, W. G., Ericsson, K. A. in The Psychology of Learning and Motivation (ed. Bower, G. H.) Vol. 16, 1–58 (Academic Press, New York, 1982).

Reif, F. Applying Cognitive Science to Education: Thinking and Learning in Scientific and Other Complex Domains (MIT Press, 2008).

Brown, A. & De Loache, J. in Siegler Children’s Thinking: What develops (L. Erlbaum Associates, 1978).

Carey, S. The origin of concepts: a précis. Behav. Brain. Sci. 34 , 113–167 (2011).

Keil, F. C. & Newman, G. in Handbook of Research on Conceptual Change (ed. Vosniadou, S.) 83–101 (Earlbaum, 2008).

Stern, E. in Pedagogy – Teaching for Learning (eds Tomlinson, P. D., Dockrell, J., Winne, P.) 153–169 (British Psychological Society, 2005).

Schneider, M. & Stern, E. The developmental relations between conceptual and procedural knowledge: a multimethod approach. Dev. Psychol. 46 (1), 178–192 (2010).

Atkinson, R. K. & Renkl, A. Interactive example-based learning environments: using interactive elements to encourage effective processing of worked examples. Educ. Psychol. Rev. 19 , 375–386 (2007).

Article   Google Scholar  

Schwartz, S., Chase, D. L., Oppezzo, C. C., M., A. & Chin, D. B. Practicing versus inventing with contrasting cases: the effects of telling first on learning and transfer. J. Educ. Psychol. 103 (4), 759–775 (2011).

Ziegler, E. & Stern, E. Delayed benefits of learning elementary algebraic transformations through contrasted comparisons. Learn. Instr. 33 , 131–146 (2014).

Zepeda, C. D., Richey, J. E., Ronevich, P. & Nokes-Malach, T. J. Direct instruction of metacognition benefits adolescent science learning, transfer, and motivation: an in vivo study. J. Educ. Psychol. 107 , 954 –970 (2015).

Anderson, L. W., Krathwohl, D. R., et al . (eds) A Taxonomy for Learning, Teaching, and Assessing: A Revision of Bloom’s Taxonomy of Educational Objectives (Allyn & Bacon, 2001).

Karmiloff-Smith, A. Beyond Modularity: A Developmental Perspective on Cognitive Science (MIT, 1992).

Spelke, E. S. & Kinzler, K. D. Core knowledge. Dev. Sci. 10 , 89–96 (2007).

Ferguson, B. & Waxman, S. R. What the [beep]? Six-month-olds link novel communicative signals to meaning. Cognition 146 , 185–189 (2016).

Waxman, S. R. & Goswami, U. in Early Childhood Development and Later Achievement (eds Pauen, S. & Bornstein, M.) (Cambridge University Press, 2012).

Pinker, S. The Stuff of Thought: Language as a Window into Human Nature (Viking, 2007).

Golinkoff, R. M., Ma, W., Song, L. & Hirsh-Pasek, K. Twenty-five years using the intermodal preferential looking paradigm to study language acquisition: What have we learned? Perspec. Psychol. Sci. 8 , 316–339 (2013).

McCrink, K. & Wynn, K. Large-number addition and subtraction by 9-month-old infants. Psychol. Sci. 15 , 776–81 (2004).

Lemer, C., Dehaene, S., Spelke, E. & Cohen, L. Approximate quantitiesand exact number words: dissociable systems. Neuropsychologia 41 , 1942–1958 (2003).

Sarnecka, B. W. & Carey, S. How counting represents number: what children must learn and when they learn it. Cognition 108 (3), 662–674 (2008).

Ifrah, G. The Universal History of Numbers (Wiley, 1999).

Alexander, A. Exploration mathematics: the rhetoric of discovery and the rise of infinitesimal methods. Configurations 9 (1), 1–36 (2001).

Lee, S. A., Sovrano, V. A. & Spelke, E. S. Navigation as a source of geometric knowledge: Young children’s use of length, angle, distance, and direction in a reorientation task. Cognition 123 , 144–161 (2012).

Dillon, M. R. & Spelke, E. S. Core geometry in perspective. Dev. Sci. 18 , 894–908 (2015).

Powell, B. B. Writing: Theory and History of the Technology of Civilization (Blackwell, 2009).

Ziegler, J. C. & Goswami, U. Becoming literate in different languages: similar problems, different solutions. Dev. Sci. 9 (5), 429–36 (2006).

Agrillo, C. Evidence for two numerical systems that are similar in humans and guppies. PLoS ONE 7 (2), e31923 (2012).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Cohen, A. et al . When is an adolescent an adult? Assessing cognitive control in emotional and non-emotional contexts. Psychol. Sci. Advance online publication 27 , 549–562 (2016).

Zelazo, P. D. The development of conscious control in childhood. Trends Cogn. Sci. 8 , 12–17 (2004).

DeLoache, J. S., &Ganea, P. A. in Learning and the Infant Mind (eds Woodward, A. & Needhman, A.) (Oxford University Press, 2009).

Binet, A., & Simon, T. The development of intelligence in children. Baltimore, Williams & Wilkins. (Reprinted 1973, New York: Arno Press; 1983, Salem, NH: Ayer Company). The 1973 volume includes reprints of many of Binet’s articles on testing (1916).

Stern, W. The Psychological Methods of Testing Intelligence (Warwick & York. No. 13 1914).

Yerkes, R. M., Bridges, J. W., & Hardwick, R. S. A Point Scale for Measuring Mental Ability (Warwick & York, 1915).

Burt, C. Handbook of Tests. For the Use in Schools (P. S. King & Son, London, 1923).

Spearman, C. General intelligence, objectively determined and measured. Am. J. Psychol. 15 , 201–293 (1904).

Jensen, A. R. The g Factor: The Science of Mental Ability . (Praeger, 1998).

Boring, E. G. Intelligence as the tests test It. New Republic 36 , 35–37 (1923).

Google Scholar  

Gottfredson, L. S. Why g matters: the complexity of everyday life. Intelligence 24 (1), S. 79–132 (1997).

Roth, B. et al. Intelligence and school grades: a meta-analysis. Intelligence 53 , 118–137 (2015).

Strenze, T. Intelligence and socioeconomic success: a metaanalytic review of longitudinal research. Intelligence 35 , S. 401–426 (2007).

Schmidt, F. L. & Hunter, J. General mental ability in the world of work: occupational attainment and job performance. J. Pers. Soc. Psychol. 86 , 162–173 (2004).

Deary, I. J., Whiteman, M. C., Starr, J., Whalley, L. J. & Fox, H. C. The impact of childhood intelligence on later life: Following up the Scottish Mental Surveys of 1932 and 1947. J. Pers. Soc. Psychol. 86 (1), 130–147 (2004).

Deary, I. J. The impact of childhood intelligence on later life: following up the Scottish mental surveys of 1932 and 1947. J. Pers. Soc. Psychol. 86 (1), 130–147 (2004).

Carroll, J. B. Human Cognitive Abilities: A Survey of Factor-Analytic Studies . (Cambridge University Press, 1993).

McGrew, K. Editorial: CHC theory and the human cognitive abilities project: Standing on the shoulders of the giants of psychometric intelligence research. Intelligence 37 , 1–10 (2009).

Bartholomew, D., Allerhand, M. & Deary, I. Measuring mental capacity: Thomson’s Bonds model and Spearman’s g-model compared. Intelligence 41 , 222–233 (2013).

Plomin, R., DeFries, J. C., Knopik, V. S., Neiderhiser, J. M. Behavioral Genetics , 6th edn, (Worth Publishers, 2013).

Plomin, R. & Deary, I. Genetics and intelligence differences: five special findings. Mol. Psychiatry 20 , 98–108 (2015).

Woltereck, R. Weitere experimentelle Untersuchungen über Artveränderung, speziell über das Wesen quantitativer Artunterschiede bei Daphniden]. Verhandlungen der deutschen zoologischen Gesellschaft 19 , 110–73 (1909).

Ridley, M. Nature via Nurture: Genes, Experience, and What Makes us Human . (HarperCollins Publishers, 2003).

Cattell, R. B. A culture-free intelligence test. J. Educ. Psychol. 31 , 161–179 (1940).

Cliffordson, C. & Gustafsson, J. E. Effects of age and schooling on intellectual performance: estimates obtained from analysis of continuous variation in age and length of schooling. Intelligence 36 , 143–152 (2008).

Schneider, W., Niklas, F. & Schmiedeler, S. Intellectual development from early childhood to early adulthood: The impact of early IQ differences on stability and change over time. Learn. Individ. Differ. 32 , 156–162 (2014).

Becker, M., Lüdtke, O., Trautwein, U., Köller, O. & Baumert, J. The differential effects of school tracking on psychometric intelligence: do academic-track schools make students smarter? J. Educ. Psychol. 104 , 682–699 (2012).

Turkheimer, E., Haley, A., Waldron, M., D’Onofrio, B. & Gottesman, I. Socioeconomic status modifies heritability of IQ in young children. Psychol. Sci. 14 , 623–628 (2003).

Tucker-Drob, E. M. & Bates, T. C. Large cross-national differences in gene x socioeconomic status interaction on intelligence. Psychol. Sci. 27 , 138–149 (2016).

Tucker-Drob, E. M. & Briley, D. A. Continuity of genetic and environmental influences on cognition across the life span: a meta-analysis of longitudinal twin and adoption studies. Psychol. Bull. 140 , 949–979 (2014).

Garaas, T. & Pomplun, M. Inspection time and visual–perceptual processing. Vision Res. 48 , 523–537 (2008).

Colom, R., Abad, F. J., Quiroga, M. A., Shih, P. C. & Flores-Mendoza, C. Working memory and intelligence are highly related constructs, but why? Intelligence 36 , 584–606 (2008).

Oberauer, K., Sü, H.-M., Wilhelm, O. & Wittmann, W. W. Which working memory functions predict intelligence? Intelligence 36 , 641–652 (2008).

Harrison, Z., Shipstead, R. & Engle, R. Why is working memory capacity related to matrix reasoning tasks? Mem. Cognit. 43 , 389–396 (2015).

Jung, R. E. & Haier, R. J. The Parieto-Frontal Integration Theory (P-FIT) of intelligence: Converging neuroimaging evidence. Behav. Brain Sci. 30 , 135–187 (2007).

Neubauer, A. C. & Fink, A. Intelligence and neural efficiency. Neurosci. Biobehav. Rev. 33 , 1004–1023 (2009).

Nussbaumer, D., Grabner, R. & Stern, E. Neural efficiency in working memory tasks: The impact of task demand. Intelligence 50 , S. 196–208 (2015).

Bornstein, M. H., Hahn, C. & Wolke, D. Systems and cascades in cognitive development and academic achievement. Child Dev. 84 , 154–162 (2013).

Pauen, S. Early Childhood Development and Later Outcome . (Cambridge University Press, 2012).

Brannon, E. M. & Van de Walle, G. A. The development of ordinal numerical competence in young children. Cognit. Psychol. 43 (1), 53–81 (2001).

Golinkoff, R. M. & Hirsh-Pasek, K. Baby wordsmith: from associationist to social sophisticate. Curr. Directions Psychol. Sci. 15 , 30–33 (2006).

Hambrick, D. Z. & Meinz, E. J. Limits on the predictive power of domain-specific experience and knowledge in skilled performance. Curr. Directions Psychol. Sci. 20 , 275–279 (2011).

Grabner, R., Stern, E. & Neubauer., A. Individual differences in chess expertise: a psychometric investigation. Acta. Psychologic 124 , 398–420 (2007).

Lubinski, D. & Benbow, C. P. Study of mathematically precocious youth after 35 years: uncovering antecedents for the development of math-science expertise. perspectives on. Psychol. Sci. 1 , 316–343 (2006).

Ericsson, K. A., Krampe, R. Th & Tesch-Römer, C. The role of deliberate practice in the acquisition of expert performance. Psychol. Rev. 100 , 363–406 (1993).

Hambrick, D. Z. et al. Deliberate practice: is that all it takes to become an expert? Intelligence 45 , 34–45 (2014).

Lubinski, D. & Benbow, C. Study of mathematically precocious youth after 35 years: uncovering antecedents for the development of math-science expertise. Pers. Psychol. Sci. 1 , 316–345 (2006).

Ackerman, P. L. & Rolfhus, E. L. The locus of adult intelligence: knowledge, abilities, and non-ability traits. Psychol. Aging. 14 , 314–330 (1999).

Rolfhus, E. L. & Ackerman, P. L. Assessing individual differences in knowledge: Knowledge structures and traits. J. Educ. Psychol. 91 , 511–526 (1999).

Kuncel, N. R. & Hezlett, S. A. Standardized tests predict graduate students’ success. Science 315 , 1080–1081 (2007).

Frey, M. C. & Detterman, D. K. Scholastic assessment or g? the relationship between the SAT and general cognitive ability. Psychol. Sci. 15 (6), 373–398 (2004).

Download references

Acknowledgements

Competing interests.

The authors declare no conflict of interest.

Author information

Authors and affiliations.

ETH Zürich, Clausiusstrasse 59, CH-8092, Zürich, Switzerland

Elsbeth Stern

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Elsbeth Stern .

Rights and permissions

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Reprints and permissions

About this article

Cite this article.

Stern, E. Individual differences in the learning potential of human beings. npj Science Learn 2 , 2 (2017). https://doi.org/10.1038/s41539-016-0003-0

Download citation

Received : 02 May 2016

Revised : 08 November 2016

Accepted : 16 November 2016

Published : 12 January 2017

DOI : https://doi.org/10.1038/s41539-016-0003-0

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

On the promise of personalized learning for educational equity.

• Hanna Dumont
• Douglas D. Ready

npj Science of Learning (2023)

Do Individual Differences Predict Change in Cognitive Training Performance? A Latent Growth Curve Modeling Approach

• Sabrina Guye
• Carla De Simoni
• Claudia C. von Bastian

Journal of Cognitive Enhancement (2017)

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

• Advanced Search

Individual differences and personalized learning: a review and appraisal

Graduate Institute of Network Learning Technology, National Central University, No. 300, Zhongda Rd., Zhongli District, 32001, Taoyüan City, Taiwan

New Citation Alert added!

This alert has been successfully added and will be sent to:

You will be notified whenever a record that you have chosen has been cited.

To manage your alert preferences, click on the button below.

New Citation Alert!

Please log in to your account

• Publisher Site

Universal Access in the Information Society

Personalized learning allows students to pursue individual learning goals at their own pace. Such a benefit is achieved by understanding each individual’s needs. Hence, it is important to implement personalized learning to accommodate students’ individual differences. Due to such importance, research into this issue has increased over the past decade. Accordingly, this paper presents a state-of-the art review of the current research that investigates relationships between individual differences and personalized learning. The main results from past research include that: (1) learning style is a major individual difference considered in works on personalized learning; (2) current works shift to address multiple individual differences, instead of a single difference; (3) learner models are widely applied to deal with multiple individual differences in the development of personalized learning; (4) learning styles, prior knowledge, preferences and ability levels are frequently considered together, and (5) it is a current trend to consider emotion recognition in the context of personalized learning. In addition to reviewing existing empirical studies, this paper also does an appraisal to identify directions for future research.

Recommendations

Review of deep learning-based personalized learning recommendation.

Online learning has become a significant way for learners to acquire knowledge, which enables anybody to learn anywhere and anytime. However, how to choose appropriate content for learning is an interesting issue, especially when one faces massive ...

Individual differences and Information Security Awareness

The main purpose of this study was to examine the relationship between individuals' Information Security Awareness (ISA) and individual difference variables, namely age, gender, personality and risk-taking propensity. Within this study, ISA was defined ...

Personalized recommendation based on review topics

The traditional collaborative filtering algorithm is a successful recommendation technology. The core idea of this algorithm is to calculate user or item similarity based on user ratings and then to predict ratings and recommend items based on similar ...

Login options

Check if you have access through your login credentials or your institution to get full access on this article.

Full Access

• Information
• Contributors

Published in

© Springer-Verlag GmbH Germany, part of Springer Nature 2020

In-Cooperation

Springer-Verlag

Berlin, Heidelberg

Publication History

• Published: 1 November 2021
• review-article

Funding Sources

Other metrics.

• Bibliometrics
• Citations 3

Article Metrics

• 3 Total Citations View Citations
• 0 Total Downloads
• Downloads (Last 12 months) 0
• Downloads (Last 6 weeks) 0

Digital Edition

View this article in digital edition.

Share this Publication link

https://dl.acm.org/doi/abs/10.1007/s10209-020-00753-4

Share on Social Media

• 0 References

Export Citations

• Please download or close your previous search result export first before starting a new bulk export. Preview is not available. By clicking download, a status dialog will open to start the export process. The process may take a few minutes but once it finishes a file will be downloadable from your browser. You may continue to browse the DL while the export process is in progress. Download
• Download citation
• Copy citation

We are preparing your search results for download ...

We will inform you here when the file is ready.

Your file of search results citations is now ready.

Your search export query has expired. Please try again.

Individual Differences in Learning and Memory Research Paper

• To find inspiration for your paper and overcome writer’s block
• As a source of information (ensure proper referencing)
• As a template for you assignment

The very idea of learning styles refers to the fact that people tend to have their own ways of accepting, evaluating, and learning information. The theories of human learning and memory retrieval subdivide the abundance of different learning styles that people use into two main categories – implicit learning and explicit learning. Implicit learning styles are based on memorizing the pieces of information, which an individual encounters during the process of learning, and the explicit learning styles come from creating and assessing different assumptions during this process. In the following paper, the variety of learning styles will be evaluated in relation to theories of human learning and memory retrieval on the basis of the findings currently made by academic researchers.

Modern specialists on cognitive psychology in education express the opinion that the success of any particular individual in education depends on which particular approach and methods related to it one is going to utilize in the educational process (Fenn & Hambrick, 2012). The approach, which an individual may utilize in one’s studies, largely depends on his or her mental abilities. Mental abilities come from the brain’s capacity and include such important skills as the skill of memorizing (Farooq & Regnier, 2011). Thus, memory is an important characteristic of one’s potential in education.

The theories of human learning and memory retrieval, which became popular over the last few decades, explain that humans tend to have different approaches to learning styles, which is explained by their cognitive peculiarities (Weng, 2012). According to the specialists in those theories, the differences in people’s learning styles should be considered by teachers when they try to adapt the requirements of the curriculum to their students’ abilities. In case they do so, the probability of greater success among students is much greater (OMIDVAR & Bee Hoon, 2012). Therefore, it is important for educators to evaluate the learning styles of their students. For example, if it appears that an individual tends to utilize implicit learning style, teachers should offer this person studying methods based on theoretical concepts, and it appears that an individual tends to utilize explicit learning styles, he or she should be offered more practical tasks aiming to research different hypotheses on the basis of results acquired in practice (Weng, 2012).

My personal learning style is a sort of combination of implicit and explicit learning with a tendency to have more characteristics of an explicit approach. This means that I try to have my vision of any subject matter that I am studying, and then I try to find evidence showing whether I am right or not to support or confute my own hypothesis. Such a learning style appears to be effective for me as I have a strong stimulus in the process of obtaining knowledge and skills. This approach also has a positive influence on my memory characteristics.

In conclusion, it should be stated that evaluating learning styles in relation to theories of human learning and memory retrieval, it appears that all humans have their own approaches to acquiring knowledge and learning, and those who use more effective learning strategies are capable of achieving better success in education. Generally, theories of human learning and memory retrieval subdivide the variety of learning styles that people use into implicit and explicit ones. Further research of a variety of academic sources and my own experience help see that explicit learning styles seem to be more effective than implicit ones.

Farooq, M., & Regnier, J. (2011). Role of Learning Styles in the Quality of Learning at Different Levels. Informatica Economica, 15 (3), 28-45.

Fenn, K. M., & Hambrick, D. Z. (2012). Individual differences in working memory capacity and learning styles. Journal Of Experimental Psychology: General, 141 (3), 404-410.

OMIDVAR, P., & Bee Hoon, T. (2012). CULTURAL VARIATIONS IN LEARNING AND LEARNING STYLES. Turkish Journal Of Distance Education (TJDE), 13 (4), 269-286.

Weng, P. (2012). The Effect of Learning Styles on Learning Strategy. Journal Of Social Sciences, 8 (2), 230-234.

• Children's Implicit and Explicit Ethnic Group
• Observing the Implicit and Explicit Communication of Rules
• Long Term Memory and Retrieval
• Assessments for ELLs: Pros and Cons
• Learning Communities and Student Success
• Acquiring External and Internal Support for the Vision
• Philosophy of Literacy Instruction
• Total Quality Management: A Viable Solution for Zayed University
• Chicago (A-D)
• Chicago (N-B)

IvyPanda. (2022, February 3). Individual Differences in Learning and Memory. https://ivypanda.com/essays/individual-differences-in-learning-and-memory/

"Individual Differences in Learning and Memory." IvyPanda , 3 Feb. 2022, ivypanda.com/essays/individual-differences-in-learning-and-memory/.

IvyPanda . (2022) 'Individual Differences in Learning and Memory'. 3 February.

IvyPanda . 2022. "Individual Differences in Learning and Memory." February 3, 2022. https://ivypanda.com/essays/individual-differences-in-learning-and-memory/.

1. IvyPanda . "Individual Differences in Learning and Memory." February 3, 2022. https://ivypanda.com/essays/individual-differences-in-learning-and-memory/.

Bibliography

IvyPanda . "Individual Differences in Learning and Memory." February 3, 2022. https://ivypanda.com/essays/individual-differences-in-learning-and-memory/.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• NPJ Sci Learn

Individual differences in the learning potential of human beings

Elsbeth stern.

ETH Zürich, Clausiusstrasse 59, CH-8092 Zürich, Switzerland

To the best of our knowledge, the genetic foundations that guide human brain development have not changed fundamentally during the past 50,000 years. However, because of their cognitive potential, humans have changed the world tremendously in the past centuries. They have invented technical devices, institutions that regulate cooperation and competition, and symbol systems, such as script and mathematics, that serve as reasoning tools. The exceptional learning ability of humans allows newborns to adapt to the world they are born into; however, there are tremendous individual differences in learning ability among humans that become obvious in school at the latest. Cognitive psychology has developed models of memory and information processing that attempt to explain how humans learn (general perspective), while the variation among individuals (differential perspective) has been the focus of psychometric intelligence research. Although both lines of research have been proceeding independently, they increasingly converge, as both investigate the concepts of working memory and knowledge construction. This review begins with presenting state-of-the-art research on human information processing and its potential in academic learning. Then, a brief overview of the history of psychometric intelligence research is combined with presenting recent work on the role of intelligence in modern societies and on the nature-nurture debate. Finally, promising approaches to integrating the general and differential perspective will be discussed in the conclusion of this review.

Human learning and information processing

In psychology textbooks, learning is commonly understood as the long-term change in mental representations and behavior as a result of experience. 1 As shown by the four criteria, learning is more than just a temporary use of information or a singular adaption to a particular situation. Rather, learning is associated with changes in mental representations that can manifest themselves in behavioral changes. Mental and behavioral changes that result from learning must be differentiated from changes that originate from internal processes, such as maturation or illness. Learning rather occurs as an interaction with the environment and is initiated to adapt personal needs to the external world.

From an evolutionary perspective, 2 living beings are born into a world in which they are continuously expected to accomplish tasks (e.g., getting food, avoiding threats, mating) to survive as individuals and as species. The brains of all types of living beings are equipped with instincts that facilitate coping with the demands of the environment to which their species has been adapted. However, because environments are variable, brains have to be flexible enough to optimize their adaptation by building new associations between various stimuli or between stimuli and responses. In the case of classical conditioning, one stimulus signals the occurrence of another stimulus and thereby allows for the anticipation of a positive or negative consequence. In the case of operant conditioning, behavior is modified by its consequence. Human beings constantly react and adapt to their environment by learning through conditioning, frequently unconsciously. 1

However, there is more to human learning than conditioning, which to the best of our knowledge, makes us different from other species. All living beings must learn how to obtain access to food in their environment, but only human beings cook and have invented numerous ways to store and conserve their food. While many animals run faster than humans and are better climbers, the construction and use of vehicles or ladders is unique to humans. There is occasional evidence of tool use among non-human species passed on to the next generation, 3 , 4 but this does not compare to the tools humans have developed that have helped them to change the world. The transition from using stonewedges for hunting to inventing wheels, cars, and iPhones within a time period of a few thousand years is a testament to the unique mental flexibility of human beings given that, to the best of our knowledge, the genes that guide human brain development have not undergone remarkable changes during the last 50,000 years. 5 This means that as a species, humans are genetically adapted to accomplish requirements of the world as it existed at approximately 48,000 BC. What is so special about human information processing? Answers to this question are usually related to the unique resource of consciousness and symbolic reasoning abilities that are, first and foremost, practiced in language. Working from here, a remarkable number of insights on human cognition have been compiled in the past decades, which now allow for a more comprehensive view of human learning.

Human learning from a general cognitive perspective

Learning manifests itself in knowledge representations processed in memory. The encoding, storage, and retrieval of information have been modeled in the multi-store model of human memory depicted in Fig.  1 . 6 Sensory memory is the earliest stage of processing the large amount of continuously incoming information from sight, hearing, and other senses. To allow goal-directed behavior and selective attention, only a fractional amount of this information passes into the working memory, which is responsible for temporarily maintaining and manipulating information during cognitive activity. 7 , 8 Working memory allows for the control of attention and thereby enables goal-directed and conscious information processing. It is the gatekeeper to long-term memory, which is assumed to have an unlimited capacity. Here, information acquired through experience and learning can be stored in different modalities as well as in symbol systems (e.g., language, script, mathematical notation systems, pictorials, music prints).

A model of human information processing, developed together with Dr. Lennart Schalk

The multi-store model of human information processing is not a one-way street, and long-term memory is not to be considered a storage room or a hard-disk where information remains unaltered once it has been deposited. A more appropriate model of long-term memory is a self-organizing network, in which verbal terms, images, or procedures are represented as interlinked nodes with varying associative strength. 9 Working memory regulates the interaction between incoming information from sensory memory and knowledge activated from long-term memory. Very strong incoming stimuli (e.g., a loud noise or a harsh light), which may signal danger, can interrupt working memory activities. For the most part, however, working memory filters out irrelevant and distracting information to ensure that the necessary goals will be achieved undisturbed. This means that working memory is continuously selecting incoming information, aligning it with knowledge retrieved from long-term memory, and preparing responses to accomplishing requirements demanded by the environment or self-set goals. Inappropriate and unsuitable information intruding from sensory as well as from long-term memory has to be inhibited, while appropriate and suitable information from both sources has to be updated. 8 The strength with which a person pursues a particular goal has an impact on the degree of inhibitory control. In case of intentional learning, working memory guards more against irrelevant information than in the case of mind wandering. Less inhibitory control makes unplanned and unintended learning possible (i.e., incidental learning).

These working memory activities are permanently changing the knowledge represented in long-term memory by adding new nodes and by altering the associative strength between them. The different formats knowledge can be represented in are listed in Fig.  1 ; some of them are more closely related to sensory input and others to abstract symbolic representations. In cognitive psychology, learning is associated with modifications of knowledge representations that allow for better use of available working memory resources. Procedural knowledge (knowing how) enables actions and is based on a production-rule system. As a consequence of repeated practice, the associations between these production rules are strengthened and will eventually result in a coordinated series of actions that can activate each other automatically with a minimum or no amount of working memory resources. This learning process not only allows for carrying out the tasks that the procedural knowledge is tailored to perform more efficiently, but also frees working memory resources that can be used for processing additional information in parallel. 10 – 12

Meaningful learning requires the construction of declarative knowledge (knowing that), which is represented in symbol systems (language, script, mathematical, or visual-spatial representations). Learning leads to the regrouping of declarative knowledge, for instance by chunking multiple unrelated pieces of knowledge into a few meaningful units. Reproducing the orally presented number series “91119893101990” is beyond working memory capacity, unless one detects two important dates of German history: the day of the fall of the Berlin Wall: 9 November 1989 and the day of reunification: 3 October 1990. Individuals who have stored both dates and can retrieve them from long-term memory are able to chunk 14 single units into two units, thereby freeing working memory resources. Memory artists, who can reproduce dozens of orally presented numbers have built a very complex knowledge base that allows for the chunking of incoming information. 13

Learning also manifests itself in the extension of declarative knowledge using concept formation and inferential reasoning. Connecting the three concepts of “animal, produce, milk” forms a basic concept of cow. Often, concepts are hierarchically related with superordinate (e.g., animal) and subordinate (e.g., cow, wombat) ordering. This provides the basis for creating meaningful knowledge by deductive reasoning. If the only thing a person knows about a wombat is that it is an animal, she can nonetheless infer that it needs food and oxygen. Depending on individual learning histories, conceptual representations can contain great variations. A farmer’s or a veterinarian’s concept of a cow is connected to many more concepts than “animal, produce, milk” and is integrated into a broader network of animals. In most farmers’ long-term memory, “cow” might be strongly connected to “pig”, while veterinarians should have particularly strong links to other ruminants. A person’s conceptual network decisively determines the selection and representation of incoming information, and it determines the profile of expertise. For many academic fields, first and foremost in the STEM area (Science, Technology, Engineering, Mathematics), it has been demonstrated that experts and novices who use the same words may have entirely different representations of their meaning. This has been convincingly demonstrated for physics and particularly in the area of mechanics. 14 Children can be considered universal novices; 15 therefore, their everyday concepts are predominantly based on characteristic features while educated adults usually consider defining features, 16 – 18 as the example of “island” demonstrates. For younger children, it primarily refers to a warm place where one can spend ones’ holidays. In contrast, adults’ concept of island does refer to a tract of land that is completely surrounded by water but not large enough to be considered a continent.

The shift from characteristic to defining features is termed “conceptual change”, 16 and promoting this kind of learning is a major challenge for school education. Students’ understanding of central concepts in an academic subject can undergo fundamental changes (e.g., the concept of weight in physics). Younger elementary school children often agree that a pile of rice has weight, but they may also deny that an individual grain of rice has weight at all. This apparently implausible answer is understandable given that younger children consider the concepts of “weight” and “being heavy” as equivalent. As such, children tend to agree that a grain of rice has weight if it is put on an ant’s back. 16 As a consequence of their education, students usually understand that an object’s weight is determined with the assistance of scales and not necessarily by personal sensation. However, representing weight as the property of an object is still not compatible with scientific physics in the Newtonian sense by which weight is conceptualized as a relation between objects. Understanding weight in this sense requires an interrelated network of knowledge, including the concepts of force, gravity, and mass (among others).

As a result of classroom instruction, students are expected to acquire procedural and conceptual knowledge of the subjects they were taught. While procedures emerge as a function of repetition and practice, the acquisition of advanced concepts, which are consistent with state of the art science, is less straightforward. 14 , 19 To support this kind of conceptual learning, insights from cognitive learning research have been integrated into educational research and are increasingly informing classroom practice. Several instructional methods have been developed and evaluated that support students in restructuring and refining their knowledge and thereby promote appropriate conceptual understanding, including self-explanations, 20 contrasting cases, 21 , 22 and metacognitive questions. 23 Cognitive research has also informed the development of the “taxonomy of learning objects”. 24 This instrument is widely employed for curriculum development and in teacher training programs to support the alignment of content-specific learning goals, means of classroom practice, and assessment. The taxonomy acknowledges the distinction between procedural and conceptual knowledge and includes six cognitive processes (listed in Fig.  1 ) that describe how knowledge can be transformed into observable achievement.

How core knowledge innate to humans can meet with academic learning

What makes humans efficient learners, however, goes beyond general memory functions discussed so far. Similar to other living beings, humans do not enter the world as empty slates 2 but are equipped with so-called core knowledge (Fig.  1 ). Evidence for core knowledge comes from preferential looking experiments with infants who are first habituated to a particular stimulus or scenario. Then, the infant is shown a second scenario that differs from the first in a specific manner. If the time he or she looks at this stimulus exceeds the looking-time at the end of the habituation phase of the first stimulus, this suggests that the infant can discriminate between the stimuli. This paradigm helps to determine whether infants detect violations of principles that underlie the physical world, such as the solidity of objects, where an object cannot occupy the same space as another object. 25 , 26 Core knowledge, which allows privileged learning and behavioral functioning with little effort, also guides the unique human ability of symbolic communication and reasoning, first and foremost, langue learning. 27 , 28 It is uncontested that humans are born with capacities for language learning, which includes the awareness of phonological, grammatical, and social aspects of language. 4 , 29 , 30

Core knowledge can serve as a starting point for the acquisition of content knowledge that has emerged as a result of cultural development. This has been examined in detail for numerical and mathematical reasoning. Two core systems have been detected in infants. As early as at 6 months of age, infants show an ability for the approximate representations of numerical magnitude, which allow them to discriminate two magnitudes depending on their ratio. 31 At the same age, the system of precise representations of distinct individuals allows infants to keep track of changes in small sets of up to three elements. 32 Mathematical competencies emerge as a result of combining both core systems and linking them to number words provided by the respective culture. 33 The Arabic place value number system, which is now common in most parts of the world, was only developed a few 100 years ago. Only after the number “0” had made its way from India via the Arabic countries to Europe were the preconditions for developing our decimal system available. 34 The Arabic number system opened up the pathway to academic mathematics. Cultural transformations based on invented symbol systems were the key to advanced mathematics. Today’s children are expected to understand concepts within a few years of schooling that took mankind centennials to develop. Central content areas in mathematics curricula of high schools, such as calculus, were only developed less than three centuries ago. 35 Given the differences between the Arabic and the Roman number systems, children born 2000 years ago could not make use of their numerical core knowledge in the same way today’s children can.

Core knowledge about navigation is meant to guide the acquisition of geometry, an area involved in numerous academic fields. 36 , 37 The cornerstone of cultural development was the invention of writing, in which language is expressed by letters or other marks. Script is a rather recent cultural invention, going back approximately 5,000 years, whereas the human genome emerged approximately 50,000 years ago. 38 Clearly, unlike oral language, humans are not directly prepared for writing and reading. Nonetheless, today, most 6-year-old children become literate during their 1st years of schooling without experiencing major obstacles. Human beings are endowed with the many skills that contribute to the ability to write and read, such as, first and foremost, language as well as auditory and visual perception and drawing. These initially independent working resources were coopted when script was invented, and teaching children to write and read at school predominantly means supporting the development of associations among these resources. 39

Part of the core knowledge innate to humans has also been found in animals, for instance numerical knowledge and geometry, but to the best of our knowledge, no other animals have invented mathematics. 40 Only humans have been able to use core knowledge for developing higher order cognition, which serves as a precondition for culture, technology, and civilization. Additionally, the unique function of human working memory is the precondition for the integration of initially independent representational systems. However, the full potential of working memory is not in place at birth, but rather matures during childhood and undergoes changes until puberty. 41 Children under the age of two are unable to switch goals 42 and memorize symbol representations appropriately. 43

To summarize what has been discussed so far, there are two sources for the exceptional learning capacity of humans. The first is the function of working memory as a general-purpose resource that allows for holding several mental representations simultaneously for further manipulation. The second is the ancient corpus of the modularized core knowledge of space, quantities, and the physical and social world. Working memory allows for the connection of this knowledge to language, numerals, and other symbol systems, which provides the basis for reasoning and the acquisition of knowledge in academic domains, if appropriate learning opportunities are provided. Both resources are innate to human beings, but they are also sources of individual differences, as will be discussed in the following sections.

Learning potentials are not alike among humans: the differential perspective

In the early twentieth century, a pragmatic need for predicting the learning potential of individuals initiated the development of standardized tests. The Frenchman Alfred Binet, who held a degree in law, constructed problems designed to determine whether children who did not meet certain school requirements suffered from mental retardation or from behavioral disturbances. 44 He asked questions that still resemble items in today’s intelligence tests; children had to repeat simple sentences and series of digits forwards and backwards as well as define words such as “house” or “money”. They were asked in what respect a fly, an ant, a butterfly and a flea are alike, and they had to reproduce drawings from memory. William Stern, an early professor of psychology at the newly founded University of Hamburg/Germany, intended to quantify individual differences in intelligence during childhood and adolescence by developing the first formula for the intelligence quotient (IQ): 45 IQ = Mental age/chronological age*100. Mental age refers to the average test score for a particular age group; this means that a 6-year-old child would have an IQ = 133 if their test score was equivalent to the mean score achieved in the group of 8-year-olds. From adolescence on, however, the average mental age scores increasingly converge, and because of the linear increase in chronological age, the IQ would decline—a trend that obviously does not match reality.

Psychologists from the United States, specifically headed by the Harvard and later Yale professor Robert Yerkes, decided to look at a person’s score relative to other people of the same age group. The average test score was assigned to an IQ = 100 by convention, and an individual’s actual score is compared to this value in terms of a standard deviation, an approach that has been retained to this day. World War I pushed the development of non-verbal intelligence tests, which were used to select young male immigrants with poor English language skills for military service. 46 In the UK, the educational psychologist Cyril Burt promoted the use of intelligence tests for assigning students to the higher academic school tracks. 47 Charles Spearman from the University College London was among the first to focus on the correlations between test items based on verbal, numerical, or visual-spatial content. 48 The substantial correlations he found provided evidence for a general intelligence model (factor-g), which has been confirmed in the following decades by numerous studies performed throughout the world. 49

The high psychometric quality of the intelligence tests constructed in different parts of the world by scientists in the early decades of the twentieth century have influenced research ever since. In 1923, Edward Boring, a leading experimental psychologist concluded, “Intelligence is what the tests test. This is a narrow definition, but it is the only point of departure for a rigorous discussion of the tests. It would be better if the psychologists could have used some other and more technical term, since the ordinary connotation of intelligence is much broader. The damage is done, however, and no harm need result if we but remember that measurable intelligence is simply what the tests of intelligence test, until further scientific observation allows us to extend the definition.”(ref. 50 , p. 37). More than 70 years later, psychologists widely agreed on a definition for intelligence originally offered by Linda Gottfredsonin 1997: “Intelligence is a very general mental capability that, among other things, involves the ability to reason, plan, solve problems, think abstractly, comprehend complex ideas, learn quickly, and learn from experience. It is not merely book learning, a narrow academic skill, or test-taking smarts. Rather, it reflects a broader and deeper capability for comprehending our surroundings—‘catching on,’ ‘making sense’ of things, or ‘figuring out’ what to do” (ref. 51 , p. 13). This definition is in line with the substantial correlations between intelligence test scores and academic success, 52 whereas correlations with measures of outside-school success, such as income or professional status, are lower but still significant. 53 , 54 Numerous longitudinal studies have revealed that IQ is a fairly stable measure across the lifespan, which has been most convincingly demonstrated in the Lothian Birth Cohorts run in Scotland. Two groups of people born in 1921 and 1936 took a test of mental ability at school when they were 11 years old. The correlation with IQ tests taken more than 60 years later was highly significant and approached r  = .70 (ref. 55 ). The same data set also demonstrated a substantial long-term impact of intelligence on various factors of life success, among them career aspects, health, and longevity. 56

Intelligence tests scores have proven to be objective, reliable, and valid measures for predicting learning outcome and more general life success. At the same time, the numerous data sets on intelligence tests that were created all over the world also contributed to a better understanding of the underlying structure of cognitive abilities. Although a factor g could be extracted in almost all data sets, correlations between subtests varied considerably, suggesting individual differences beyond general cognitive capabilities. Modality factors (verbal, numerical, or visual spatial) have been observed, showing increased correlations between tests based on the same modality, but requiring different mental operations. On the other hand, increased correlations were also observed between tests based on different modalities, but similar mental operations (e.g., either memorizing or reasoning). The hierarchical structure of intelligence, with factor g on the top and specific factors beneath, was quite obvious from the very beginning of running statistical analyses with intelligence items. Nonetheless, it appeared a major challenge for intelligence researchers to agree on a taxonomy of abilities on the second and subsequent levels. In 1993, John Carroll published his synthesis of hundreds of published data sets on the structure of intelligence after decades of research. 57 In his suggested three-stratum model, factor g is the top layer, with the middle layer encompassing broader abilities such as comprehension knowledge, reasoning, quantitative knowledge, reading and writing, and visual and auditory processing. Eighty narrower abilities, such as spatial scanning, oral production fluency, and sound discrimination, are located in the bottom layer. To date, Carroll’s work is considered the most comprehensive view of the structure of individual variations in cognitive abilities. 58 However, the interpretation of factor g is still under discussion among scientists. Factor g could be a comprehensive characteristic of the brain that makes information processing generally more or less efficient (top-down-approach). Existing data sets, however, are also compatible with a model of intelligence according to which the human brain is comprised of a large number of single abilities that have to be sampled for mental work (bottom-up approach). In this case, factor g can be considered a statistical correlate that is an emerging synergy of narrow abilities. 59

Genetic sources of individual differences in intelligence

From studies with identical and fraternal twins, we know that genetic differences can explain a considerable amount of variance in IQ. The correlation between test scores of identical twins raised together approaches r  = .80 and thereby is almost equal to the reliability coefficient of the respective test. On the other hand, IQ-correlations between raised-together same-sex fraternal twins are rarely higher than .50, a value also found for regular siblings. Given that the shared environment for regular siblings is lower than for fraternal twins, this result qualifies the impact of environmental factors on intelligence. The amount of genetic variance is judged in statistical analyses based on the difference between the intra-pair correlations for identical and fraternal twins. 60 High rates of heritability, however, do not mean that we can gauge a person’s cognitive capabilities from his or her DNA. The search for the genes responsible for the expression of cognitive capabilities has not yet had much success, despite the money and effort invested in human genome projects. It is entirely plausible that intelligence is formed by a very large number of genes, each with a small effect, spread out across the entire genome. Moreover, these genes seem to interact in very complicated ways with each other as well as with environmental cues. 61

An entirely false but nonetheless still widespread misunderstanding is to equate “genetic sources” with “inevitability” because people fail to recognize the existence of reaction norms, a concept invented in 1909 by the German biologist, Richard Woltereck. Reaction norms depict the range of phenotypes a genotype can produce depending on the environment. 62 For some few physiological individual characteristics (e.g., the color of eyes) the reaction norm is quite narrow, which means gene expression will rarely be affected by varying environments. Other physiological characteristics, such as height, have a high degree of heritability and a large reaction norm. Whether an individual reaches the height made possible by the genome depends on the nutrition during childhood and adolescence. In a wealthy country with uniform access to food, average height will be larger than in a poor country with many malnourished inhabitants. However, within both countries, people vary in height. The heritability in the wealthy country can be expected to approach 100% because everybody enjoyed sufficient nutrition. In contrast, in the poor country, some were sufficiently nourished and, therefore, reached the height expressed by their genome, while others were malnourished and, therefore, remained smaller than their genes would have allowed under more favorable conditions. For height, the reaction norm is quite large because gene expression depends on nutrition during childhood and adolescence. This explains the well-documented tendency for people who have grown up in developed countries to become progressively taller in the past decades.

The environment regulates gene expression, which means that instead of “nature vs. nurture”, a more accurate phrase is “nature via nurture”. 63 The complex interaction between genes and environment can also explain the fact that heritability of intelligence increases during the lifespan. 61 This well-established finding is a result of societies in which a broad variety of cognitive activities available in professional and private life enable adults more than children to actively select special environments that fit their genes. People who have found their niche can perfect their competencies by deliberate learning.

In the first decades of developing intelligence tests, researchers were naive to the validity of non-verbal intelligence; so-called culture-free or culture-fair tests, based on visual-spatial material such as mirror images, mazes or series and matrices of geometric figures, were supposed to be suitable for studying people of different social and cultural levels. 64 This is now considered incorrect because in the meantime, there has been overwhelming evidence for the impact of schooling on the development of intelligence and the establishment and stabilization of individual differences. Approximately 10 years of institutionalized education is necessary for the intelligence of individuals to approach its maximum potential. 65 – 67

Altogether, twin and adoption studies suggest that 50–80% of IQ variation is due to genetic differences. 61 This relatively large range in the percentage across different studies is due to the heritability of intelligence in the population studied, specifically, the large reaction norm of the genes giving rise to the development of intelligence. Generally, the amount of variance in intelligence test scores explained by genes is higher the more society members have access to school education, health care, and sufficient nutrition. There is strong evidence for a decrease in the heritability of intelligence for children from families with lower socioeconomic status (SES). For example, lower SES fraternal twins resembled each other more than higher SES ones, indicating a stronger impact of shared environment under the former condition. 68 In other words, because of the less stimulating environment in lower SES families, the expression of genes involved in the development of intelligence is likely to be hampered. Although it may be counterintuitive at first, this suggests that a high heritability rate of intelligence in a society is an indicator of economic and educational equity. Additionally, this means that countries that ensure access to nutrition, health care, and high quality education independent of social background enable their members to develop their intelligence according to their genetic potential. This was confirmed by a meta-analysis on interactions between SES and heritability rate. While studies run in the United States showed a positive correlation between SES and heritability rate, studies from Western Europe countries and Australia with a higher degree of economic and social equality did not. 69 , 70

Cognitive processes behind intelligence test scores: how individuals differ in information processing

In the first part of this paper, cognitive processes were discussed that, in principle, enable human beings to develop the academic competencies that are particularly advantageous in our world today. In the second part, intelligence test scores were shown to be valid indicators of academic and professional success, and differences in IQ were shown to have sound genetic sources. Over many decades, research on cognitive processes and psychometric intelligence has been developing largely independently of one another, but in the meantime, they have converged. Tests that were developed to provide evidence for the different components of human cognition revealed large individual differences and were substantially correlated with intelligence tests. Tests of memory function were correlated with tests of factor g. Sensory memory tests have shown that the exposure duration required for reliably identifying a simple stimulus (inspection time) is negatively correlatedwith intelligence. 71 For working memory, there is a large body of research indicating substantial relationships between all types of working memory functions and IQ, with average correlations >.50 (refs 72 – 74 ). In these studies, working memory functions are measured by speed tasks that require goal-oriented active monitoring of incoming information or reactions under interfering and distracting conditions. Neural efficiency has been identified as a major neural characteristic of intelligence; more intelligent individuals show less brain activation (measured by electroencephalogram or functional magnetic resonance imaging) when completing intelligence test items 75 , 76 as well as working memory items. 77 Differences in information-processing efficiency were already found in 4-month-old children. Most importantly, they could predict psychometric intelligence in 8-year-old children. 78

These results clearly suggest that a portion of individual differences can be traced back to differences in domain-general cognitive competencies. However, psychometric research also shows that individual differences do exist beyond factor g on a more specific level. Differences in numerical, language, and spatial abilities are well established. Longitudinal studies starting in infancy suggest that sources of these differences may be traced back to variations in core knowledge. Non-symbolic numerical competencies in infancy have an impact on mathematical achievement. 79 Similar long-term effects were found for other areas of core knowledge, 80 particularly language. 81

Endowed with general and specific cognitive resources, human beings growing up in modern societies are exposed to informal and formal learning environments that foster the acquisition of procedural as well as declarative knowledge in areas that are part of the school curriculum. Being endowed with genes that support efficient working memory functions and that provide the basis for usable core knowledge allows for the exploitation of learning opportunities provided by the environment. This facilitates the acquisition of knowledge that is broad as well as deep enough to be prepared for mastering the, as of yet, unknown demands of the future. 18 Regression analyses based on longitudinal studies have revealed that the confounded variance of prior knowledge and intelligence predicts learning outcome and expertise better than each single variable. 82 – 84 Importantly, no matter how intelligent a person is, gaining expertise in a complex and sophisticated field requires deliberate practice and an immense investment of time. 85 However, intelligence differences will come into play in the amount of time that has to be invested to reach a certain degree of expertise. 86 Moreover, intelligence builds a barrier to content areas in which a person can excel. As discussed in the first part of this paper, some content areas—first and foremost from STEM fields—are characterized by abstract concepts mainly based on defining features, which are themselves integrated into a broader network of other abstract concepts and procedures. Only individuals who clearly score above average on intelligence tests can excel in these areas. 84 , 87 For individuals who were fortunate enough to attend schools that offered high-quality education, intelligence and measures of deep and broad knowledge are highly correlated. 88 , 89 A strong impact of general intelligence has also been shown for university entrance tests such as the SAT, which mainly ask for the application of knowledge in new fields. 90 , 91 Societies that provide uniform access to cognitively stimulating environments help individuals to achieve their potential but also bring to bear differences in intelligence. Education is not the great equalizer, but rather generates individual differences rooted in genes.

Acknowledgements

Competing interests.

The authors declare no conflict of interest.

Individual Differences in Second Language Learning: the Road Ahead

• Published: 07 October 2021
• Volume 45 , pages 237–244, ( 2021 )

Cite this article

• Wen-Ta Tseng 1 &
• Xuesong (Andy) Gao 2  

6744 Accesses

5 Citations

1 Altmetric

Explore all metrics

Avoid common mistakes on your manuscript.

Introduction

In the past few years, several special issues on individual differences (IDs) in language learning have been published in the field of language education [ 6 , 17 ], Oga-Baldwin, Fryer, & Larson-Hall [ 24 ]. These publications range from a broad scope of theme collections to a specific focus of topic modeling. For instance, the contributions to the special issue edited by de Bot and Bátyi feature a wide range of IDs constructs including age, aptitude, attitude, and motivation. On the other hand, the contributions to a more recent special issue edited by Oga-Baldwin et al. focus on language learning motivation as modeled and interpreted through a diverse lens of mainstream motivational theories such as expectancy-value theory, self-determination theory, personal investment theory, and goal theories. A much more recent special issue edited by Gurzynski-Weiss places a predominant emphasis on the investigation of the dynamic nature of individual differences in L2 learning. The empirical studies featured in these three special issues not only point to the ongoing and enduring commitment collectively embraced by IDs researchers, but also underline the challenges concerning whether IDs in language learning should be structuralized or systematized with reference to exclusively one particular theoretical framework or dual conceptually distinct yet implicitly complementary theoretical underpinnings. To elaborate, unlike IDs constructs such as anxiety, aptitude, or willingness to communicate, which receive relatively few conceptual challenges, research on both language learning motivation and language learning strategies (LLS) has undergone different stages of transitions regarding the theoretical frameworks behind the two IDs constructs [ 1 , 29 ]

Language Learning Motivation

The major theoretical transition in L2 motivation is featured by moving from operationalizing L2 motivation via a social-educational modeling approach to adopting a self-based viewpoint to modeling L2 motivation (e.g., [ 31 ]). The central construct underlying the social-educational model refers to integrativeness or integrative motivation , whereas the pivotal factor underlying the L2 self model refers to ideal L2 self . Despite attempts over the years to reinterpret or even replace integrativeness with ideal L2 self, as theorized by Dörnyei and his associates, it is important to note that the two conceptually distinct L2 motivational models receive equal credits from academia. Over the years, the fundamental discrepancy regarding theoretical underpinnings between the two L2 motivational models has become even clearer. As Gardner [ 16 ] critically remarks:

Cognition and affect are parallel systems. One is not superior to the other… The L2 self is a cognitive model while that of integrative motivation is an affective one… Their utility is in the validity of the models, not in their superiority over others (p. 226).

A decade later, Dörnyei [ 13 ] also explicitly recognized the theoretical divide between the two L2 motivation models:

[T]he type of identification adopted in the L2 Motivational Self System – identification with a projected future image within the person’s self-concept, rather than identification with an external reference group such as the L2 community as was the case with the notion integrativeness – can serve certain purposes (p. xx).

After nearly 30 years of numerous rounds of theoretical debates and empirical testing, readers interested in L2 motivation research can finally obtain a clear picture that integrative motivation and ideal L2 self are in actuality isomorphic and complementary. This is especially true concerning the pragmatic valences in describing and explaining learners’ motivated language learning behaviors. Integrative motivation and ideal L2 self also differ in essence both in origin and by target, leading Claro (2020) to suggest that “the ideal L2 self cannot replace integrativeness” (p. 253). Dörnyei [ 13 ] also expects to see a wave of “renewed vibrancy” (p. xxi) in bringing integrative motivation back to the spotlight. As expected, one of the aims of the present Special Issue is to respond to the urgent call to action that has yet to be answered in the two prior IDs special issues.

Language Learning Strategies

The other aim of the present Special Issue is to address a similar research controversy which has remained in the field of LLS for more than 15 years. Interestingly, analogous to the debate over the concept of integrativeness and integrative motivation, the criticisms levelled against LLS are mainly twofold: one centers around the definition of strategies , and the other revolves around the validity of the rating scale structure underlying Oxford’s [ 28 ] Strategy Inventory of Language Learning (SILL) [ 8 , 10 , 11 , 12 , 32 ]. As with the case of L2 motivation research, Dörnyei and Skehan argued that the operational definitions of learning strategies theorized by Oxford [ 27 , 28 ] and O’Malley and Chamot [ 25 , 26 ] tended to be “inconsistent and elusive” (p. 608). Dörnyei [ 12 ] took a bold step to equate learning strategies to “idiosyncratic self-regulated behaviour” (p. 183). Critically, Dörnyei never shows any substantial empirical evidence to prove learning strategies are idiosyncratic at all, and nor does Dörnyei ever theorize why strategies should be conceptualized as behavior rather than as technique or influences in the way he defines motivational strategies [ 9 ]. For instance, Dörnyei ([ 9 ], p. 28) defines motivational strategies in the same paragraph as “techniques that promote the individual’s goal-related behaviours” and “ those motivational influences that are consciously exerted to achieve some systematic and enduring positive effect [ emphasis original],” respectively. Following these two definitions, it is clear that Dörnyei equated strategies to techniques and influences, but not to behavior. It is by no means clear and consistent as to why techniques can be equivalent to influences in an operational sense. Following this line of thinking, it seems fair to say that the unjustified criticisms such as “inconsistent and illusive” thrown onto learning strategies may become equally relevant and valid to motivational strategies. In this way, should motivational strategies be likewise considered idiosyncratic as in the case of learning strategies? Furthermore, Hadfield & Dörnyei [ 19 ] created the term achievement strategies to refer to “study techniques that can be used across a range of tasks to improve learning” (p. 146), and on the same page further argued that it was imperative that learners be introduced to the “techniques that might help them to work more productively, getting them to discuss and evaluate these and finally selecting those that work best for them” (p. 146). Notably, it is clear that Dörnyei also formally associates strategies with the construct of learning achievement, which by definition is equivalent to the concept of learning strategies. When taken together, a careful review of Dörnyei’s work on motivational strategies published between 2001 and 2015 seems to show that Dörnyei has been forbidding others to do what he is doing himself. Two sets of standards might have been applied to learning strategies and motivational strategies respectively, the constructs of which are, to a certain extent, interrelated.

Dörnyei [ 10 , 11 ] further introduced the term self-regulation , a term which he argues is more capable of reflecting the concept of strategic learning. In particular, he proposed a five-factor model to indicate the possible underlying construct of self-regulatory capacity of language learning. The proposed five-factor measurement model was initially sent into empirical testing in English vocabulary learning [ 32 , 33 ] as an attempt to complement SILL, which is in principle operationalized by behavioral items. Indeed, the call for the paradigm shift has received enormous attention from the field and raised practitioners and researchers’ awareness of the divide between the quality and quantity dimensions of strategic learning. Over the years, it has been observed that the coexistence of strategy use (the quantity dimension) and proactive control of strategy use (the quality dimension) have greatly advanced readers’ understanding of the underlying theoretical underpinnings of strategic learning. The empirical findings of gender differences have shed light on how the two complementary forms of strategic learning may become integrated to support brain study. The effect of gender differences on strategy use started to draw researchers’ attention with the rise of SILL in the 1990s [ 14 , 18 , 30 ]. Notably, in their very large-scale empirical study ( N  > 1200), Oxford and Nyikos [ 30 ] noted that females reported more frequent strategy use than males on a latent factor called formal rule-related practice strategies . This strategic factor greatly capitalized on learners’ cognitive ability to analyze and understand the linguistic codes and rules of a target language. In another study, Ehrman & Oxford [ 14 ] further found that females also reported more frequent strategy use than males on metacognitive strategic behaviors such as checking, monitoring, and planning one’s learning performance. Importantly, the findings of significant gender differences in both cognitive and metacognitive strategy use in these early primary studies provide indirect yet critical support for the later findings uncovered by functional magnetic resonance imaging (fMRI) techniques [ 2 , 20 , 22 ]. Overall, females showed a stronger and wider response than males not only in the amygdala, which is responsible for emotional regulation, but also in the prefrontal cortex areas, where cognitive processing and higher order mental functioning such as planning, monitoring, and problem solving occur [ 3 ]. The above discussion suggests that the quantity dimension of strategic learning as operationalized by SILL can offer diagnostic information regarding gender differences in cognitive and metacognitive functioning in the task of language learning.

On the other hand, Tseng, Liu, and Nix [ 34 ] developed and validated an instrument to tap into the quality dimension of strategy use (i.e., the proactive control of strategy use) in language learning: Self-Regulatory Control Scale for Language Learning (SR lang ). Unlike what Dörnyei [ 10 ] has hypothesized, a four-factor measurement model consisting of boredom control , awareness control , goal control , and emotion control was procured. Importantly, the naming of the four factors was essentially theories-referenced, rather than intuition-guided, which allows for the inferencing of gender differences in the proactive control of strategy use in language learning. In their second phase of validity study, Tseng et al. employed latent regression modeling to further check the way in which gender difference would modulate learners’ control of strategy use over the four dimensions in SR lang . Results indicated that females had significantly stronger and better proactive control of strategy use than males in boredom control, awareness control, and emotion control, but not in goal control. Essentially, because boredom control and emotion control are intrinsically associated with emotion regulation [ 4 , 21 ], the neurobiological function of which the amygdala is responsible for. Sensibly, therefore, Tseng et al.’s findings can be credited with being significantly, indirectly notwithstanding, convergent with those by fMRI [ 2 , 20 , 22 ], in which the responsive magnitude in emotional regulation was directly observed in the brain and more active in females than in males. Critically, the foregoing discussion further suggests that the quality dimension of strategic learning as operationalized by SR lang may provide useful information concerning gender differences in emotional regulation in the task of language learning. In sum, the recent research findings based on fMRI studies clearly suggest that both the frequency component and the control component of strategy use can be considered complementary regarding their theoretical value in depicting and explicating gender differences in strategic learning of a language.

Introducing the Special Issue

The orthodoxical stance of learning strategies, as well as the unique empirical significance of the two complementary forms of strategy use, has yet to be articulated and showcased in prior special issues. To address the research gap, therefore, the second aim of the present Special Issue is to inform and update readers in a timely manner of the academic merits which have been historically built-in, but have yet to receive sufficient justice from the field. In total, the current Special Issue features 7 articles which offer a balanced report and insightful update of empirical research in relation to L2 motivation and LLS research. The first section includes three articles addressing the research on L2 motivation. The first article (Kim & Shin) examined the mediating role of integrative motivation in the causal link between self-efficacy and English achievement in a Korean sample. Through the aid of the bootstrapping technique, Kim and Shin found that the mediating effect exerted by integrative motivation was significant and meaningful. Kim and Shin’s research findings suggest that learners’ affective identification with the L2 community group works in synergy with learners’ cognitive belief of how well they can achieve in studying a foreign language.

The second article (Cheng) in the first section investigated the effects of grit and L2 self on willingness to communicate (WTC) in a Taiwanese sample. In Cheng’s study, the scale targeting grit—passion of and persistence toward a specific ultimate goal—was conceptualized by two dimensions: consistency of interest (COI) and persistence of effort (POE). The instrument measuring L2 self was operationalized by four factors: Ideal L2 self own , Ideal L2 self other , Ought-to L2 self own , Ought-to L2 self other . By using hierarchical regression modeling, Cheng found that three types of L2 self-images (ideal L2 self own , ideal L2 self other , and Ought-to L2 self own ), taken together with grit, could jointly exert explanatory power over WTC. Cheng suggests that both establishing a gritty attitude and shaping an ideal L2 self vision carry equal weight in sustaining L2 motivation.

In the third article of the first section, Soltanian and Ghapanchi approached L2 motivation from the viewpoint of “investment,” a concept that views L2 learning as an entity of social practice. Soltanian and Ghapanchi’s study explored the factors that might affect Iranian EFL learners’ investment through a qualitative inquiry. The results of their study revealed that the economic, social, cultural, and symbolic capital jointly influenced varying degrees of Iranian EFL learners’ willingness to learn English. The reason for including a paper which focused on investment is that both investment theory [ 7 , 23 ] and Gardner’s [ 15 ] socio-education model underlined language learners’ connections to the social world they live in.

The second section of the present Special Issue contains the other three articles with a focus on strategic learning. The first article (Nathan et al.) was featured by constructing a systematic review of the core components of LLS research conducted in Taiwan. Upon an extensive, thorough search of literature via numerous databases, Nathan et al. successfully extracted 100 empirical studies eligible for systematic review. The 100 primary studies were analyzed based on three evaluation criteria: (a) contexts and participant characteristics; (b) theoretical-conceptual aspects; and (c) methodological characteristics. The researchers observed that, as a whole, there was a pendulum shift from a predominant adoption of survey tools to a more diversified deployment of multiple research approaches and recognized the shift as positive. Nathan et al. suggest that situating LLS in a social-politically unique context (i.e., Taiwan) helps move the research on LLS ahead in an even more global context.

In the second article of the second section, Haga and Reinders investigated the emotional regulation of feedback on language learning in a sample of diverse L1 backgrounds including Bulgarian, Hungarian, Mexican, Polish, and Russian. Haga and Reinders applied dynamic systems theory (DST) to record and systematize a large set of interview data collected from 25 participants. Their findings showed that although participants experienced a wide range of positive and negative emotions, negative emotions could have facilitative effects on language learning, particularly on shaping learners’ multilingual identities. Haga and Reinders suggested that future research needs to be oriented to a deeper understanding of emotions and emotional regulation in language learning.

The third study (Koenig & Guertler) included in the second section conducted a two-phase large-scale consecutive survey study to explore German learners’ thoughts and perceptions of improvement and satisfaction regarding their self-regulated language learning. Sample 1 and sample 2 involved 1646 and 796 participants situated in the German higher educational context. The results of the phase I survey study showed that time investment in self-regulated study could lead to greater language skill improvement. The results of the phase II retrospective survey study further indicated that German college learners had not yet cultivated enough required capacity to self-regulate their language learning strategies. The findings of Koenig and Guertler’s study pointed to the individuality and variations observed at the nexus where the quality dimension (self-regulatory capacity) and the quantity dimension (use of language learning strategies) of strategic learning intersected.

The present Special Issue concludes with an updated critical review of the role of individual differences in language learning and teaching from a complex-dynamic and socio-ecological perspective. This wide lens allows readers to holistically visualize the road ahead of IDs research in language learning. Griffiths featured 11 salient IDs factors considered to be important in L2 classrooms. Based on the results of an empirical investigation with L2 teachers, Griffiths pointed out that motivation and strategy use were ranked as the most important factors and the other factors such as aptitude and gender as at least somewhat important. The findings led Griffiths to suggest that there is a need for future research to take a holistic approach to advance the understanding of the potential interplay among the 11 salient IDs factors in language learning.

To conclude, we believe the 7 articles collected in this Special Issue have made unique yet valuable contributions to the field of IDs research in language learning. Upon the publication of the Special Issue, we hope that researchers and practitioners alike can be enlightened in a timely manner by the theoretical clarifications critically needed in the field. Readers should look forward to not only renewed, but also heightened vibrancy regarding innovative applications and integrations of different IDs factors deemed to be significant and critical in the field of language learning.

Al-Hoorie, A. H., & MacIntyre, P. D. (Eds.). (2020). Contemporary language motivation theory: 60 years since Gardner and Lambert (1959) . Multilingual Matters.

Google Scholar  

AlRyalat, S. A. (2017). Gender similarities and differences in brain activation strategies: Voxel-based meta-analysis on fMRI studies. Journal of Integrative Neuroscience, 16 , 227–240.

Article   Google Scholar  

Baars, B., & Gage, N. M. (2010). Cognition, brain, and consciousness: Introduction to cognitive neuroscience (2nd ed.). Elsevier Ltd.

Boekaerts, M. (2011). Emotions, emotion regulation, and self-regulation of learning. In B. J. Zimmerman & D. J. Schunk (Eds.), Handbook of self-regulation of learning and performance (pp. 408–425). Routledge.

Claro, J. (2019). Identification with External and Internal Referents: Integrativeness and the Ideal L2 Self. In A. Al-Hoorie & P. MacIntyre (Eds.), Contemporary Language Motivation Theory (pp. 233–261). Multilingual Matters.

de Bot, K., & Bátyi, S. (Eds.). (2017). Individual differences in second language development [Special issue]. Studies in Second Language Learning and Teaching , 7(1).

Darvin, R., & Norton, B. (2015). Identity and a model of investment in applied linguistics. Annual Review of Applied Linguistics, 35 , 36–56.

Dörnyei, Z., & Skehan, P. (2003). Individual differences in second language learning. In C. J. Doughty & M. H. Long (Eds.), The handbook of second language acquisition (pp. 589–630). Blackwell.

Chapter   Google Scholar  

Dörnyei, Z. (2001). Motivational strategies in the language classroom . Cambridge University Press.

Book   Google Scholar  

Dörnyei, Z. (2005). The psychology of the language learner: Individual differences in second language acquisition . Lawrence Erlbaum.

Dörnyei, Z. (2006). Individual differences in second language acquisition. AILA Review, 19 , 42–68.

Dörnyei, Z. (2009). The psychology of second language acquisition . Oxford University Press.

Dörnyei, Z. (2020). Foreward. In A. Al-Hoorie & P. MacIntyre (Eds.), Contemporary language motivation theory (pp. xix–xxi). Multilingual Matters.

Ehrman, M., & Oxford, R. (1989). Effects of sex differences, career choice, and psychological type on adult language learning strategies. The Modern Language Journal, 73 , 1–13.

Gardner, R. C. (1985). Social psychology and second language learning: The role of attitudes and motivation . Edward Arnold.

Gardner, R. C. (2010). Motivation and second language acquisition . Peter Lang Publishing Inc.

Gurzynski-Weiss, L. (2020). Investigating the dynamic nature of learner individual differences in L2 learning. [Special issue]. Studies in Second Language Learning and Teaching , 10 (1).

Green, J. M., & Oxford, R. (1995). A closer look at learning strategies, L2 proficiency, and gender. TESOL Quarterly, 29 , 261–297.

Hadfield, J., & Dörnyei, Z. (2013). Motivating learning . Longman.

Koch, K., Pauly, K., Kellermann, T., Seiferth, N. Y., Reske, M., Backes, V., & Habel, U. (2007). Gender differences in the cognitive control of emotion: An fMRI study. Neuropsychologia, 45 , 2744–2754.

Macklem, G. L. (2015). Boredom in the classroom: Addressing student motivation, self- regulation and engagement in learning . Springer.

McRae, K., Ochsner, K. N., Mauss, I. B., Gabrieli, J. D., & Gross, J. J. (2008). Gender differences in emotion regulation: An fMRI study of cognitive reappraisal. Group Processes & Intergroup Relations, 11 , 143–162.

Norton, B. (2013).  Identity and language learning : Extending the conversation. Bristol: Multilingual matters.

Oga-Baldwin, W. L. Q., Fryer, L. K., & Larson-Hall, J. (2019). The critical role of the individual in language education: New directions from the learning sciences [Special issue].  System ,  86 .

O’Malley, J. M., & Chamot, A. U. (1990). Learning strategies in second language acquisition . Cambridge University Press.

O’Malley, J. M., & Chamot, A. U. (1994). Learning strategies in second language learning. The International Encyclopedia of Education (Vol. 6, pp. 3329–3335). Pergamon Press.

Oxford, R. L. (1989). Use of language learning strategies: A synthesis of studies with implications for strategy training. System, 17 , 235–247.

Oxford, R. L. (1990). Language learning strategies: What every teacher should know . Heinle and Heinle.

Oxford, R. L. (2017). Teaching and researching language learning strategies: Self-regulation in context (2nd ed.). Taylor & Francis.

Oxford, R., & Nyikos, M. (1989). Variables affecting choice of language learning strategies by university students. The Modern Language Journal, 73 , 291–300.

Schmidtke-Bode, K., & Kachel, G. (2020). Exploring the motivational antecedents of Nepalese learners of L2 English. International Review of Applied Linguistics in Language Teaching, 58 (4), 379–418.

Tseng, W.-T., Dörnyei, Z., & Schmitt, N. (2006). A new approach to assessing strategic learning: The case of self-regulation in vocabulary acquisition. Applied Linguistics, 27 , 78–102.

Tseng, W. T., & Schmitt, N. (2008). Toward a model of motivated vocabulary learning: A structural equation modeling approach. Language Learning, 58 , 357–400.

Tseng, W. T., Liu, H., & Nix, J. M. L. (2017). Self-regulation in language learning: Scale validation and gender effects. Perceptual and Motor Skills, 124 , 531–548.

Download references

Author information

Authors and affiliations.

National Taiwan University of Science & Technology, Taipei City, Taiwan

Wen-Ta Tseng

University of New South Wales, Sydney, Australia

Xuesong (Andy) Gao

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Wen-Ta Tseng .

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Tseng, WT., Gao, X.(. Individual Differences in Second Language Learning: the Road Ahead. English Teaching & Learning 45 , 237–244 (2021). https://doi.org/10.1007/s42321-021-00097-z

Download citation

Published : 07 October 2021

Issue Date : October 2021

DOI : https://doi.org/10.1007/s42321-021-00097-z

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Find a journal
• Publish with us
• Track your research

Thursday, September 6, 2012

Reflection paper - individual differences, no comments:, post a comment.

What is generative AI?

In the months and years since ChatGPT burst on the scene in November 2022, generative AI (gen AI) has come a long way. Every month sees the launch of new tools, rules, or iterative technological advancements. While many have reacted to ChatGPT (and AI and machine learning more broadly) with fear, machine learning clearly has the potential for good. In the years since its wide deployment, machine learning has demonstrated impact in a number of industries, accomplishing things like medical imaging analysis  and high-resolution weather forecasts. A 2022 McKinsey survey shows that AI adoption has more than doubled  over the past five years, and investment in AI is increasing apace. It’s clear that generative AI tools like ChatGPT (the GPT stands for generative pretrained transformer) and image generator DALL-E (its name a mashup of the surrealist artist Salvador Dalí and the lovable Pixar robot WALL-E) have the potential to change how a range of jobs are performed. The full scope of that impact, though, is still unknown—as are the risks.

Get to know and directly engage with McKinsey's senior experts on generative AI

Aamer Baig is a senior partner in McKinsey’s Chicago office;  Lareina Yee  is a senior partner in the Bay Area office; and senior partners  Alex Singla  and Alexander Sukharevsky , global leaders of QuantumBlack, AI by McKinsey, are based in the Chicago and London offices, respectively.

Still, organizations of all stripes have raced to incorporate gen AI tools into their business models, looking to capture a piece of a sizable prize. McKinsey research indicates that gen AI applications stand to add up to $4.4 trillion  to the global economy—annually. Indeed, it seems possible that within the next three years, anything in the technology, media, and telecommunications space not connected to AI will be considered obsolete or ineffective .

But before all that value can be raked in, we need to get a few things straight: What is gen AI, how was it developed, and what does it mean for people and organizations? Read on to get the download.

To stay up to date on this critical topic, sign up for email alerts on “artificial intelligence” here .

Learn more about QuantumBlack , AI by McKinsey.

What every CEO should know about generative AI

What’s the difference between machine learning and artificial intelligence, about quantumblack, ai by mckinsey.

QuantumBlack, McKinsey’s AI arm, helps companies transform using the power of technology, technical expertise, and industry experts. With thousands of practitioners at QuantumBlack (data engineers, data scientists, product managers, designers, and software engineers) and McKinsey (industry and domain experts), we are working to solve the world’s most important AI challenges. QuantumBlack Labs is our center of technology development and client innovation, which has been driving cutting-edge advancements and developments in AI through locations across the globe.

Artificial intelligence is pretty much just what it sounds like—the practice of getting machines to mimic human intelligence to perform tasks. You’ve probably interacted with AI even if you don’t realize it—voice assistants like Siri and Alexa are founded on AI technology, as are customer service chatbots that pop up to help you navigate websites.

Machine learning is a type of artificial intelligence. Through machine learning, practitioners develop artificial intelligence through models that can “learn” from data patterns without human direction. The unmanageably huge volume and complexity of data (unmanageable by humans, anyway) that is now being generated has increased machine learning’s potential , as well as the need for it.

What are the main types of machine learning models?

Machine learning is founded on a number of building blocks, starting with classical statistical techniques  developed between the 18th and 20th centuries for small data sets. In the 1930s and 1940s, the pioneers of computing—including theoretical mathematician Alan Turing—began working on the basic techniques for machine learning. But these techniques were limited to laboratories until the late 1970s, when scientists first developed computers powerful enough to mount them.

Until recently, machine learning was largely limited to predictive models, used to observe and classify patterns in content. For example, a classic machine learning problem is to start with an image or several images of, say, adorable cats. The program would then identify patterns among the images, and then scrutinize random images for ones that would match the adorable cat pattern. Generative AI was a breakthrough. Rather than simply perceive and classify a photo of a cat, machine learning is now able to create an image or text description of a cat on demand.

Introducing McKinsey Explainers : Direct answers to complex questions

How do text-based machine learning models work how are they trained.

ChatGPT may be getting all the headlines now, but it’s not the first text-based machine learning model to make a splash. OpenAI’s GPT-3 and Google’s BERT both launched in recent years to some fanfare. But before ChatGPT, which by most accounts works pretty well most of the time (though it’s still being evaluated), AI chatbots didn’t always get the best reviews. GPT-3 is “by turns super impressive and super disappointing,” said New York Times tech reporter Cade Metz in a video where he and food writer Priya Krishna asked GPT-3 to write recipes for a (rather disastrous) Thanksgiving dinner .

The first machine learning models to work with text were trained by humans to classify various inputs according to labels set by researchers. One example would be a model trained to label social media  posts as either positive or negative. This type of training is known as supervised learning because a human is in charge of “teaching” the model what to do.

The next generation of text-based machine learning models rely on what’s known as self-supervised learning. This type of training involves feeding a model a massive amount of text so it becomes able to generate predictions. For example, some models can predict, based on a few words, how a sentence will end. With the right amount of sample text—say, a broad swath of the internet—these text models become quite accurate. We’re seeing just how accurate with the success of tools like ChatGPT.

What does it take to build a generative AI model?

Building a generative AI model has for the most part been a major undertaking, to the extent that only a few well-resourced tech heavyweights have made an attempt . OpenAI, the company behind ChatGPT, former GPT models, and DALL-E, has billions in funding from bold-face-name donors. DeepMind is a subsidiary of Alphabet, the parent company of Google, and even Meta has dipped a toe into the generative AI model pool with its Make-A-Video product. These companies employ some of the world’s best computer scientists and engineers.

But it’s not just talent. When you’re asking a model to train using nearly the entire internet, it’s going to cost you. OpenAI hasn’t released exact costs, but estimates indicate that GPT-3 was trained on around 45 terabytes of text data—that’s about one million feet of bookshelf space, or a quarter of the entire Library of Congress—at an estimated cost of several million dollars. These aren’t resources your garden-variety start-up can access.

What kinds of output can a generative AI model produce?

As you may have noticed above, outputs from generative AI models can be indistinguishable from human-generated content, or they can seem a little uncanny. The results depend on the quality of the model—as we’ve seen, ChatGPT’s outputs so far appear superior to those of its predecessors—and the match between the model and the use case, or input.

ChatGPT can produce what one commentator called a “ solid A- ” essay comparing theories of nationalism from Benedict Anderson and Ernest Gellner—in ten seconds. It also produced an already famous passage describing how to remove a peanut butter sandwich from a VCR in the style of the King James Bible. Image-generating AI models like DALL-E 2 can create strange, beautiful images on demand, like a Raphael painting of a Madonna and child, eating pizza . Other generative AI models can produce code, video, audio, or business simulations .

But the outputs aren’t always accurate—or appropriate. When Priya Krishna asked DALL-E 2 to come up with an image for Thanksgiving dinner, it produced a scene where the turkey was garnished with whole limes, set next to a bowl of what appeared to be guacamole. For its part, ChatGPT seems to have trouble counting, or solving basic algebra problems—or, indeed, overcoming the sexist and racist bias that lurks in the undercurrents of the internet and society more broadly.

Generative AI outputs are carefully calibrated combinations of the data used to train the algorithms. Because the amount of data used to train these algorithms is so incredibly massive—as noted, GPT-3 was trained on 45 terabytes of text data—the models can appear to be “creative” when producing outputs. What’s more, the models usually have random elements, which means they can produce a variety of outputs from one input request—making them seem even more lifelike.

What kinds of problems can a generative AI model solve?

The opportunity for businesses is clear. Generative AI tools can produce a wide variety of credible writing in seconds, then respond to criticism to make the writing more fit for purpose. This has implications for a wide variety of industries, from IT and software organizations that can benefit from the instantaneous, largely correct code generated by AI models to organizations in need of marketing copy. In short, any organization that needs to produce clear written materials potentially stands to benefit. Organizations can also use generative AI to create more technical materials, such as higher-resolution versions of medical images. And with the time and resources saved here, organizations can pursue new business opportunities and the chance to create more value.

We’ve seen that developing a generative AI model is so resource intensive that it is out of the question for all but the biggest and best-resourced companies. Companies looking to put generative AI to work have the option to either use generative AI out of the box or fine-tune them to perform a specific task. If you need to prepare slides according to a specific style, for example, you could ask the model to “learn” how headlines are normally written based on the data in the slides, then feed it slide data and ask it to write appropriate headlines.

What are the limitations of AI models? How can these potentially be overcome?

Because they are so new, we have yet to see the long tail effect of generative AI models. This means there are some inherent risks  involved in using them—some known and some unknown.

The outputs generative AI models produce may often sound extremely convincing. This is by design. But sometimes the information they generate is just plain wrong. Worse, sometimes it’s biased (because it’s built on the gender, racial, and myriad other biases of the internet and society more generally) and can be manipulated to enable unethical or criminal activity. For example, ChatGPT won’t give you instructions on how to hotwire a car, but if you say you need to hotwire a car to save a baby, the algorithm is happy to comply. Organizations that rely on generative AI models should reckon with reputational and legal risks involved in unintentionally publishing biased, offensive, or copyrighted content.

These risks can be mitigated, however, in a few ways. For one, it’s crucial to carefully select the initial data used to train these models to avoid including toxic or biased content. Next, rather than employing an off-the-shelf generative AI model, organizations could consider using smaller, specialized models. Organizations with more resources could also customize a general model based on their own data to fit their needs and minimize biases. Organizations should also keep a human in the loop (that is, to make sure a real human checks the output of a generative AI model before it is published or used) and avoid using generative AI models for critical decisions, such as those involving significant resources or human welfare.

It can’t be emphasized enough that this is a new field. The landscape of risks and opportunities  is likely to change rapidly in coming weeks, months, and years. New use cases are being tested monthly, and new models are likely to be developed in the coming years. As generative AI becomes increasingly, and seamlessly, incorporated into business, society, and our personal lives, we can also expect a new regulatory climate  to take shape. As organizations begin experimenting—and creating value—with these tools, leaders will do well to keep a finger on the pulse of regulation and risk.

Articles referenced include:

• " Implementing generative AI with speed and safety ,” March 13, 2024, Oliver Bevan, Michael Chui , Ida Kristensen , Brittany Presten, and Lareina Yee
• “ Beyond the hype: Capturing the potential of AI and gen AI in tech, media, and telecom ,” February 22, 2024, Venkat Atluri , Peter Dahlström , Brendan Gaffey , Víctor García de la Torre, Noshir Kaka , Tomás Lajous , Alex Singla , Alex Sukharevsky , Andrea Travasoni , and Benjamim Vieira
• “ As gen AI advances, regulators—and risk functions—rush to keep pace ,” December 21, 2023, Andreas Kremer, Angela Luget, Daniel Mikkelsen , Henning Soller , Malin Strandell-Jansson, and Sheila Zingg
• “ The economic potential of generative AI: The next productivity frontier ,” June 14, 2023, Michael Chui , Eric Hazan , Roger Roberts , Alex Singla , Kate Smaje , Alex Sukharevsky , Lareina Yee , and Rodney Zemmel
• “ What every CEO should know about generative AI ,” May 12, 2023, Michael Chui , Roger Roberts , Tanya Rodchenko, Alex Singla , Alex Sukharevsky , Lareina Yee , and Delphine Zurkiya
• “ Exploring opportunities in the generative AI value chain ,” April 26, 2023, Tobias Härlin, Gardar Björnsson Rova , Alex Singla , Oleg Sokolov, and Alex Sukharevsky
• “ The state of AI in 2022—and a half decade in review ,” December 6, 2022,  Michael Chui ,  Bryce Hall ,  Helen Mayhew , Alex Singla , and Alex Sukharevsky
• “ McKinsey Technology Trends Outlook 2023 ,” July 20, 2023,  Michael Chui , Mena Issler,  Roger Roberts , and  Lareina Yee  
• “ An executive’s guide to AI ,” Michael Chui , Vishnu Kamalnath, and Brian McCarthy
• “ What AI can and can’t do (yet) for your business ,” January 11, 2018,  Michael Chui , James Manyika , and Mehdi Miremadi

This article was updated in April 2024; it was originally published in January 2023.

Want to know more about generative AI?

Related articles.

The data dividend: Fueling generative AI

Don’t wait—create, with generative AI