Key Skills for the 21st Century-an evidence-based review-part-4.pdf

KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 18 Education: Future Frontiers | Analytical Report visible in behaviours only (even though they can be measured indirectly in self-rating questionnaires). Finally, further analysis is needed to determine the extent to which these...

KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 18 Education: Future Frontiers | Analytical Report visible in behaviours only (even though they can be measured indirectly in self-rating questionnaires). Finally, further analysis is needed to determine the extent to which these theoretical constructs are helpful for teachers and systems to use and whether they result in improved student learning and development. Key skills for the 21st century The reason for presenting the organising frameworks was to offer an overview from different sources of the range of skills seen as important for education in the 21st century. All focus on student learning and progress, and propose comprehensive sets of skills that in�uence young people’s future success. The models recognise the importance of cognitive skills, but also emphasise the role of other sorts of skills such as those associated with social and emotional learning and student dispositions towards and engagement in learning. The different sets of skills included in the frameworks form models of relationships representing the mechanisms by which student skills and dispositions shape student learning and future success. Source: Derived from Lamb et al., 2015 Figure 2-5 International Study of City Youth Framework for 21st Century Skills KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 19 Education: Future Frontiers | Analytical Report While the frameworks are useful as a preliminary classi�cation of the complex phenomenon understood under the concept of 21st century skills, it is also important to explore individual skills in depth, in order to identify the evidence supporting their importance for schools. Therefore, attention now turns to the individual skills that the literature identi�es as vital for young people to acquire in order to do well in school, work and the world more broadly. For each skill, existing literature was reviewed to clarify the meaning of the construct, its relationship with academic performance, its malleability in educational contexts and its possible transferability. It is important to note at the outset that more work is needed to gain a comprehensive understanding of each skill. The case-by-case overview below presents just a brief outline of current research on the most commonly cited skills. Nine skills or constructs are presented:1. critical thinking 2. creativity 3. metacognition4. problem solving 5. collaboration 6. motivation 7. self-ef�cacy 8. conscientiousness, and 9. grit or perseverance. This list, while not exhaustive, highlights the skills that have received the most attention in recent educational policy and research. It is also important to note that this section does not provide a systematic analysis of the complex relationships existing between the different constructs, though the importance of this is acknowledged. Rather, relationships between skills are considered on a case- by-case basis, when it is important for making sense of linked constructs and their educational relevance. It is also worth noting that there are important skills not analysed in this report. The teaching of literacy and numeracy skills, for example, is well established and accepted as a legitimate and critical part of the goals of schooling. So too it is assumed that quality schooling encompasses an appropriate breadth and depth of core subjects. For this reason, they are not discussed in this report. ICT-related competencies, de�ned as the mastery of various technology- based environments and tools, are also now stated learning objectives in Australian schooling, and their formalisation into teaching, learning and assessment materials and practices is becoming well established. An in-depth study of this and digital literacy more broadly, whilst acknowledged as important components of many 21st century frameworks, is considered outside of the scope of this current investigation. Critical thinking Critical thinking as a concept has a long tradition of research in different �elds. Critical thinking research primarily involves two academic communities: cognitive psychologists on the one hand (Halpern, 1998) and philosophers on the other (Ennis, 1991) 1. These communities have contrasted conceptions of critical thinking. While cognitive psychologists tend to emphasise the cognitive processes and ways of thinking that de�ne critical thinking, philosophers tend to outline the ideal dispositions and attributes of a critical thinker. Despite their disagreements, researchers from different traditions agree that critical thinking entails a judgement or evaluation for analysing claims, arguments and evidence and for making inferences using deductive and inductive reasoning to solve a problem or make a decision (Lai, 2011; Lai & Viering, 2012, p. 12). Critical thinking as a skill refers to the ability to assess the value of a claim or information and come to a conclusion about what to believe or to do about it. This could be taken as a generic de�nition of critical thinking. While the psychological stream of research has received more attention than the philosophical one, KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 20 Education: Future Frontiers | Analytical Report including in educational debates, critical thinking researchers from all horizons recognise that cognitive processes and modes of thought are insuf�cient to exhibit critical thinking. Dispositions (or ‘habits of mind’) are an integral part of critical thinking as well (Facione et al., 1990; Facione, et al., 1995). The most commonly mentioned dispositions include open- mindedness (Bailin et al., 1999b), inquisitiveness (Facione et al., 2000), the desire to seek information (Ennis, 1985) and a willingness to consider the point of view of others (Facione et al., 1990). As with all the skills discussed in this report, the question of the generality or domain speci�city of the skill divides researchers. This is a major tension in the research on skills in general. Nevertheless, critical thinking researchers tend to agree that a level of background knowledge is imperative for thinking critically (Bailin et al., 1999a; McPeck, 1990; Toplak & Stanovich, 2002; Willingham, 2008). To think critically, students need to apply thinking on something for which they need certain content knowledge, and a lack of content knowledge hinders the expression of critical thinking skills. There is also the matter of context dependence. Students may be skilled at thinking critically in mathematics while failing to do so in English or science (Lai & Viering, 2012, p. 44). Yet, there is no consensus on the extent to which critical thinking is context-dependent (or domain speci�c), partly because of the different meanings given to domain speci�city by different researchers. This lack of agreement on the meaning of a ‘general’ or ‘speci�c’ skill is another framing line of the entire �eld of research on skills. The lack of agreement on the degree of domain speci�city for critical thinking implies, by de�nition, that researchers do not agree on the extent to which critical thinking is a transferable skill. Little research exists that examines the conditions of transferability of critical thinking across tasks within a subject or across subjects within the curriculum (not to mention transfer beyond the classroom and the education system). De�nitional and measurement issues make this type of research dif�cult to conduct. The malleability of critical thinking is less debated than its context speci�city. Researchers have accumulated reliable evidence—including a meta- analysis of previous research—showing not only that critical thinking can be developed and learned, but also that it can be learned at school, i.e. instructional interventions support the development of critical thinking skills (Abrami et al., 2008; Kennedy et al., 1991; Lai, 2011). In other words, based on the way critical thinking is commonly de�ned and measured in the literature, critical thinking is a teachable skill. There is little research investigating whether critical thinking is a correlate to higher student achievement in school or to longer term outcomes linked to work, social integration and citizenship. Tentative results have been found for an association between critical thinking skills and achievement in a narrow range of contexts in higher education ( Gadzella et al., 1997; Williams et al., 2003), but more research is needed at different levels of education and in other contexts to obtain more conclusive results. Nevertheless, critical thinking is, along with possibly metacognition, one of the most well-researched skills. Creativity Creativity is often associated with critical thinking in discussions on skills. The Australian Curriculum, for example, acknowledges the strong links between them through the general capability of ‘critical and creative thinking’. In some respects, this is understandable, since critical thinking is often seen as a condition for creativity and vice versa. At the conceptual level, however, critical thinking and creativity can be distinguished in a meaningful way. The same absence of consensus over the meaning of the construct is as evident in research on creativity as it is on critical thinking. There is no agreed-upon de�nition of creativity that most researchers use, even though most argue that it entails the production of something recognised as novel or useful in a given social context (Plucker et al., 2004). The fact that an output must be socially recognised as valuable KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 21 Education: Future Frontiers | Analytical Report (in addition to being original) to be considered as creative highlights that (1) creativity is a skill sitting at the intersection between the individual and society, and (2) creative skills can be restricted to speci�c social contexts. The social ‘situatedness’ of creativity explains the dif�culty researchers have in addressing the question of transferability. Since accepting a product or performance as creative rests on a social judgement, the degree of originality needed to label something as creative cannot be decided a priori. Another similarity with critical thinking is that creativity is generally seen as requiring more than technical skills (Sternberg, 2006a). Important dispositions and related skills underpin students’ creative abilities, such as motivation, the ability to take risks, open-mindedness to new ideas and a capacity to tolerate ambiguity (Sternberg, 2010). Creativity is also seen as closely related with other cognitive skills such as problem identi�cation, idea generation, and problem solving. As is common with most skills mentioned in this review, the question of the context-dependence of creativity is highly debated (Baer, 2016; Barbot et al., 2016). A wide range of theories of creativity exist (e.g. developmental, economic, psychometric, cognitive, problem-based, evolutionary, typological) with different approaches to the meaning of creativity and its context dependence (Kozbelt et al., 2010). Some of these approaches highlight the importance of social and cultural contexts for creativity (Lubart, 2010) and the context speci�city of creativity even within a given domain (e.g. art) (Lubart & Guignard, 2004), while others argue that creativity is both a general and domain-speci�c ability (Milgram & Livne, 2005; Plucker, 2005; Sternberg, 2005). Drawing on these arguments, it appears that creativity depends on a familiarity with contextual knowledge, but the extent to which it is possible to identify generic components to creativity is uncertain. At the very least, current empirical research suggests that creativity cannot be considered as generic only (Han & Marvin, 2002). Creativity has received particular attention from researchers interested in gifted and talented education ( Guignard et al., 2016; Runco, 1993). While the extent to which creativity can be developed by all students in educational contexts is unclear, it is generally believed that creativity is not a skill reserved to a small minority of students only. Developing creative skills is accessible to all students when adequate didactical and pedagogical conditions exist. Anecdotal evidence also suggests that deliberate interventions can improve student creativity in educational contexts (Maker, 2004). Interestingly, creativity has been found to be a good predictor of future achievement (in higher education), even after accounting for past academic results (Sternberg, 2006b). As with most critical skills, however, the results are overwhelmingly correlational in nature: there is no agreement on the causal relationship at play between creativity and student success. Metacognition The term ‘metacognition’ was used by Flavell to describe thinking about an individual’s cognitive processes and activity (i.e. thinking about thinking, or meta-thinking) (Flavell, 1979). This form of cognitive self-management (Kuhn & Dean, 2004) is a complex skill comprising both cognitive self-knowledge and active cognitive self-monitoring (Schraw et al., 2006; Schraw & Moshman, 1995). Metacognition is often subsumed under the broader term of ‘self-regulated learning’, encompassing ‘the set of intrapersonal processes by which individuals remain on course in their pursuit of goals they have adopted’ (Hoyle & Davisson, 2011, p. 6). As a dimension of self-regulated learning, metacognition is associated with improved learning and academic performance (Ford et al., 1998; Pintrich & de Groot, 1990; Winne & Nesbit, 2010; Zimmerman, 1990). Yet, the literature still faces issues for making sense of the �rmly established correlation existing between self-regulated learning and academic achievement, especially since the emergence of a consensus on KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 22 Education: Future Frontiers | Analytical Report an adequate instrument for assessing self-regulated learning is yet to come (Lennon, 2010). One of the main ways in which self-regulated learning in general (and thus metacognition in particular) appears to in�uence academic achievement is through students’ sense of agency (Hacker et al., 2009). The question of the domain speci�city of metacognitive skills is a complex debate extending beyond the pure question of de�nition and measurement. The domain speci�city of skills such as metacognition is a theoretical and empirical question that shapes the educational use of psychological research and evidence (Dunlosky et al., 2009). While it is possible to develop a relatively general level of metacognitive knowledge, metacognitive monitoring is shaped by the task at hand or the problem to be solved. As with all other skills, the domain speci�city of metacognition (and self-regulated learning skills in general) is uncertain (Tobias & Everson, 2009). Psychologists have progressively come to emphasise the social aspects of self-regulation (Alexander, 1995; Pressley, 1995), and the contextual factors shaping metacognitive regulation are to be taken into consideration. Metacognition can be taught and metacognitive skills tend to develop over the school years (Bryce & Whitebread, 2012; Schneider, 2008, 2015; Veenman, 2015; Weil et al., 2013). An analysis of research on metacognitive teaching strategies, published in 2014, reported a range of effective strategies for promoting metacognition for both primary and secondary school students, concluding that “metacognitive strategies are applicable across different disciplines and grade levels and they are effective for teaching both content knowledge and academic skills.” Instructional practices most frequently used included teacher modelling with Think Aloud, diagramming, practice, answer checking, checklists, and goal attainment (Ellis et al., 2014, p. 4021). A similar study looking at the application to science teaching reported that among effective classroom strategies were enquiry-based learning, the role of collaborative support, strategy and problem solving instruction, and the construction of mental models. These instructional strategies were identi�ed, according to the authors, because they re�ected extensive research over the previous decade within the science education literature and were essential to metacognition and self-regulation (Schraw et al., 2006). Metacognition is seen as involving both knowledge about cognitive processes and strategies for monitoring these processes (Serra & Metcalfe, 2009). The development of student metacognition is best engaged in speci�c curriculum areas, since metacognitive skills depend on content knowledge and expertise (Bransford et al., 2000). Teaching metacognitive skills can be organised systematically, and researchers recommend doing so within the context of subject areas (Bransford et al., 2000, p. 21). Adopting a metacognitive approach to student learning in school subjects can help them in taking ownership of their own learning and developing the sense of agency mentioned above (Bransford et al., 2000, p. 18; Hacker et al., 2009). Problem solving Problem solving is a core skill relevant to most academic activities, but also to most tasks in the workplace. Problem solving is traditionally seen as having three main components involving: (1) the selection of strategies to solve a given problem, (2) the application of strategies to this problem, and (3) the monitoring of the strategies used to solve the problem (Newell, 1990). Metacognition (the third component) is thus an integral part of problem solving. In most conventional measures of cognitive ability, complex problem solving is one of the skills being tested. It is thus logical that (complex) problem solving is associated with cognitive ability ( Greiff & Neubert, 2014; Lotz et al., 2016; Stadler et al., 2015). In the context of education, the OECD has de�ned problem solving as follows ( OECD, 2014, p. 30): ‘An individual’s capacity to engage in cognitive processing to understand and resolve problem situations where a method of solution is not immediately obvious. It includes the willingness KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 23 Education: Future Frontiers | Analytical Report to engage with such situations in order to achieve one’s potential as a constructive and re�ective citizen’ Based on its review of 21st century skills, the US National Research Council places a particular emphasis on problem solving and metacognition as part of their overview of transferable skills (Pellegrino & Hilton, 2012, p. 169). Both are cognitive skills that have a demonstrated relationship with improved educational outcomes, although this relationship is mainly expressed via an improvement in cognitive ability. For example, a recent study showed that problem solving is not correlated with academic achievement once cognitive ability is accounted for, except in mathematics (Lotz et al., 2016). However, the extent to which problem solving is associated with academic performance after accounting for cognitive ability depends on the de�nition and measurement of cognitive ability and problem solving ( Greiff & Neubert, 2014; Stadler et al., 2015). Generally speaking, problem-solving skills seem to have broad ranging relevance in academic contexts and beyond. Problem solving is one of the key skills for which the possibility of transfer is most promising. Problem solving skills can be transferable when students understand the ‘underlying principles of what was learned’ when encountering the initial problem (Pellegrino & Hilton, 2012, p. 98). However, it is important to note that such transfer is only likely to occur when faced with structurally comparable problems, often within a given subject. Another important condition is to teach speci�cally for the transfer of problem solving skills rather than assuming that the transfer will take place automatically (Phye, 2001). Student understanding of the ‘underlying principles’ behind a problem and its solution(s) can be supported through instruction, as is the mental representation of a problem and the ability to navigate different representations of a problem (all elements of transferable problem solving skills). This provides supporting evidence that problem solving skills are malleable and susceptible to improvement within educational contexts. Structuring student learning around problems to be solved can be bene�cial for a range of cognitive skills. Learning activities structured around problems to solve or projects to conduct seem to be prime vehicles for the acquisition of transferable cognitive skills (Pellegrino & Hilton, 2012, p. 147). Ideally, such problems should be authentic and grounded in scenarios common to daily life, but these types of learning activities are much more dif�cult to assess in a standardised manner. Promising recent developments have emerged in collaborative problem solving, based on the premise that these skills are most useful in the types of problems encountered by workers in Western economies. The OECD’s focus on collaborative problem solving in PISA 2015 ( OECD, 2017) is an interesting initiative. It is also worth mentioning that researchers have developed a framework to teach collaborative problem solving that breaks the construct into social and cognitive skills that students need in equal measure (Hesse et al., 2015). Problem solving may provide an avenue through which education systems could support more collective forms of learning and assessment, requiring interpersonal skills such as cooperation. Collaboration and cooperation In some respects, interpersonal skills such as collaboration and cooperation are the most contentious skills amongst those discussed in this report. Collaboration is often conceived of as a social skill, alongside assertiveness, responsibility and empathy (Malecki & Elliott, 2002). In addition to customary de�nitional issues about the meaning of ‘collaboration’ and ‘cooperation’, ‘there are few well- established practical assessments for interpersonal competencies that are suitable for use in schools’ (Pellegrino & Hilton, 2012, p. 148). The dif�culty in measuring collaboration in schools (Webb, 1995, 1997) makes its systematic inclusion into classroom practices problematic. KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 24 Education: Future Frontiers | Analytical Report Most education is structured around individual learning and assessment, and the role of collaboration and cooperation is only recognised at the margins of individual student learning. Compared to most other skills, social skills such as collaboration, empathy or responsibility tend to have a weak correlation with student grades (Farrington et al., 2012, p. 11). No clear consensus emerges from the literature on social skills and achievement across all stages of education. Most of the research on the relationship between social skills and academic performance has been conducted at the primary school level. Longitudinal research suggests that social skills can predict achievement in primary and secondary education (Teo et al., 1996). A more recent meta-analysis asserts that social skills are associated with academic learning in schools (Durlak et al., 2011). However, in their review of social skills, Farrington et al. (2012, p. 48) outline the dif�culty of drawing any conclusions from the literature on the relationship between social skills and academic achievement since most studies ‘confound social skills with other variables’. In fact, most correlational studies provide little information as to the direction of the relationship between constructs. However, it would be a mistake to expect a technical answer to a political or social question. Just as correlations between individual skills and academic achievement are insuf�cient to turn these skills into legitimate learning outcomes, the lack of association between interpersonal skills and academic achievement is not suf�cient for regarding the skills as legitimate learning outcomes. Indeed, the lack of association is largely a consequence of the way academic achievement is currently de�ned and measured, i.e. as the individual mastery of knowledge and skills in speci�c disciplines or areas of knowledge. In fact, certain collaborative practices in the classroom can foster student learning (Bossert, 1988). While the recent OECD initiative of including collaborative problem solving in PISA 2015 is an original endeavour, inferring individual competency from group performance remains problematic (Dijkstra et al., 2016; Frykedal & Chiriac, 2011; Webb, 1993), even with the use of information technology. Motivation Motivation is a �eld of research with a longer history than most other skills. It is often de�ned as the impetus to engage in purposive behaviour (Ryan & Deci, 2000). The literature has come to distinguish intrinsic motivation, where individuals are moved by personal interests and desires, from extrinsic motivation, where individuals’ purposive behaviour is driven by external rewards or sanctions. Although the distinction between intrinsic and extrinsic motivation is not always clear-cut, this dichotomy offers the advantage of facilitating empirical research on interventions aimed at enhancing motivation. In short, motivation is based on speci�c interests, preferences, and perceptions that drive individuals to engage (or not engage) in an activity. Motivation researchers have progressively come to uncover a ‘greater complexity of motivational processes and multiple levels of in�uence’ than previously envisaged (Wentzel & Wig�eld, 2009, p. 2). Motivation is a multifaceted construct aggregating beliefs, values, goals and needs (Wentzel & Wig�eld, 2009). The growing recognition of social and contextual in�uences on student motivation has also resulted in a more elaborate conception of motivation. Moreover, motivation is related in complex ways with other dispositions and skills. The next section shows that self-ef�cacy and locus of control are associated with student motivation. Similarly, interests (Hidi & Harackiewicz, 2000; Hidi & Renninger, 2006) and goals (Broussard & Garrison, 2004) shape student motivation for academic tasks and thus indirectly in�uence their educational chances. This increasingly sophisticated picture of student motivation has made the task of conducting applied research on ‘best practice’ in school settings more dif�cult. Categorising motivation as a skill is somewhat questionable, since identifying motivation as a form of developing expertise raises conceptual KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 25 Education: Future Frontiers | Analytical Report challenges. Based on the International Study of City Youth framework, it is perhaps best described as a disposition or mindset expressed through behaviours and engagement. In any case, researchers have found a stable association between motivation and performance in mathematics and reading, especially in elementary school (Broussard & Garrison, 2004; Gottfried, 1990). Motivation is also associated with transferable learning (Yeager & Walton, 2011): students who are motivated by an activity, problem or subjects are more likely to develop transferable knowledge and skills (at least within a certain area of transfer) in this activity. However, this must not be interpreted as meaning that motivation itself is a generally transferable skill. In fact, in the �eld of motivation research, domain speci�city is more �rmly established than it is in research on critical thinking or creativity. Moreover, the domain speci�city of motivation is not identical for all students: empirical research suggests not only that academic (intrinsic) motivation tends to decrease with age, but also that particular subjects in�uence this trend, with motivation declining at a greater rate in mathematics than in social studies ( Gottfried et al., 2001). This con�rms that the structures of the curriculum and the educational conditions in which students are placed can affect their level of motivation. One general implication from this �nding is that academic motivation can be taught and learned, and a range of effective interventions have proven successful at fostering student motivation in schools (Wig�eld & Wentzel, 2007). A recent meta-analysis of academic motivation interventions found them to be generally effective (Lazowski & Hulleman, 2016). Self-e�cacy and locus of control (sense of agency) As suggested above, motivation is shaped by perception of self and the task or problem at hand. In this respect, self-ef�cacy, de�ned as perceived ability to succeed, as well as sense of agency (i.e. locus of control), de�ned as belief that you are in control of the outcome of the activity, underpin motivation (Bandura, 1982, 1993; Eccles & Wig�eld, 2002; Pintrich & de Groot, 1990). This is a clear illustration of the complex web of relationships existing between various dispositions, attitudes and skills. At a conceptual level, this complex web of relationships may also allude to the existence of mechanisms of circular causation at play between various skills which are generally measured independently. Self-ef�cacy and locus of control are often studied together in the literature, as they both refer to an individual’s sense of control over the outcome of a task or activity. As noted above, self-ef�cacy can be de�ned as a belief in one’s own ability to do or complete something and can be expressed with the statement ‘I can do it’ while locus of control is the sense of in�uence an individual feels over things and can be expressed with the statement ‘Doing well is up to me, rather than others’. Both of these mindsets have been found to be consistently associated with student outcomes (Bandura, 1997; Cury et al., 2006; Pajares, 1996; Pajares & Graham, 1999). The two constructs, sometimes labelled as ‘academic mindsets’ (Farrington et al., 2012, p. 10), have been shown to in�uence school success (as measured by student grades). They also have another feature in common: their in�uence on academic performance is primarily indirect, that is, via their impact on behavioural expressions measured by other constructs such as academic perseverance and motivation. Interestingly, student perceptions of the nature of cognitive ability are particularly important for shaping student behaviours. Students who see success as a product of effort are more likely to engage and persevere in academic endeavours as opposed to those who see it more as a product of ‘innate’ ability (Yeager & Walton, 2011). While tentative evidence suggests that educational interventions can improve self-ef�cacy and sense of agency, further research is needed to broaden the evidence base (Yeager & Walton, 2011). Speci�c instructional and pedagogical approaches can support self-ef�cacy and locus of control mindsets, at least within the context of a given activity. This implies that the constructs are malleable and students’ KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 26 Education: Future Frontiers | Analytical Report self-ef�cacy and sense of agency are susceptible to change. The malleability of these academic mindsets has been demonstrated in experimental contexts, and contextual factors in more natural settings (e.g. a classroom) also shape students’ sense of agency and self-ef�cacy (Farrington et al., 2012, p. 32). As mentioned above, supporting the development of metacognitive skills and practices is likely to enhance self-ef�cacy and locus of control beliefs. One of the more uncertain aspects of research on these two constructs is the question of transfer. As the analysis of all the skills so far demonstrates, the question of the transferability is the most contentious area of research. As of yet, researchers seem to know little about the individual contextual and task-related conditions of transferability of skills. Conscientiousness As a �rst approximation, conscientiousness can be de�ned as a form of self-discipline. Conscientiousness is expressed as a diligent behaviour based on self-control and application to a given problem, task or activity. It is a multi-faceted skill, with some researchers identifying three facets: industriousness, impulse-control and orderliness ( Costantini & Perugini, 2016) while others identify as many as eight dimensions (Rikoon et al., 2016). Conscientiousness entertains complex relationships with various other skills, including motivation, locus of control and, above all, tenacity or grit. Conscientiousness is considered by personality psychologists as one of the big �ve personality traits, alongside openness to experience, extraversion, agreeableness and neuroticism. In differential psychology, these personality traits are seen as relatively stable over time (Matthews et al., 2009). They are often considered as useful signposts to characterising an individual’s personality. While ‘relatively stable’ may be of little help in understanding the malleability of personality traits, a recent review of available evidence suggests that personality traits are at least ‘not set in stone’ (Almlund et al., 2011, p. 9) and can be shaped by educational experiences or other interventions. Interestingly, conscientiousness is the only one of the ‘big �ve’ personality traits that shows a consistent association with performance in school and higher education (Farrington et al., 2012, p. 24; Richardson & Abraham, 2009; Tackman et al., 2017). Recent meta-analyses of previous research even suggest that the effect size for the association between conscientiousness and academic performance is comparable to the effect size for the association between cognitive ability (as measured with IQ tests) and academic performance (Poropat, 2009, 2014). Conscientiousness is also the only personality trait correlated with work performance across a range of performance measures (Barrick et al., 2001). Finally, amongst the competencies classi�ed as ‘non-cognitive’ by the US Committee on Deeper Learning, conscientiousness (staying organised and being committed to study or work) is the one most highly correlated with desirable educational and occupational outcomes (Pellegrino & Hilton, 2012, p. 4). Although the availability of causal evidence remains limited, conscientiousness seems to have a signi�cant in�uence on achievement, above and beyond IQ (Duckworth & Seligman, 2005). In schools, this can take the form of ‘academic tenacity’ (Dweck et al., 2014). Although the importance of such academic mindsets for achievement seems well established, the evidence available for the ability to develop these skills in the classroom is scant. Conscientiousness, like other personality traits, is a complex construct shaped by a range of developmental in�uences (Srivastava et al., 2003) and associated with a range of social factors (Furnham & Cheng, 2014). Evidence suggests that personality traits can change over time. They also depend on life experiences (Roberts & DelVecchio, 2000). Importantly for educational practice, persistence in a given activity or task can be supported by speci�c KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 27 Education: Future Frontiers | Analytical Report interventions and the shaping of the environment. This bring us to the notions of ‘perseverance’ and ‘grit’. Grit and perseverance Perseverance can be conceptualised as a dimension of conscientiousness. In an academic context, grit can be de�ned as commitment and perseverance in learning tasks and activities (long-term goals) despite dif�culties (or obstacles). Academic perseverance or tenacity generally relies on goal-setting and accepting delayed grati�cation (Farrington et al., 2012, p. 9). Answering the question of the malleability of perseverance depends signi�cantly on the de�nition used. While the ‘relatively stable’ nature of conscientiousness, de�ned as a personality trait, would suggest that educational interventions aimed at improving student conscientiousness are fruitless if they focus on academic perseverance, speci�cally, research shows that this disposition or skill is malleable. As Farrington et al. summarise it: ‘There is signi�cant empirical evidence that students demonstrate different amounts of perseverance at academic tasks under differing conditions, supporting the idea that academic perseverance as a behaviour in a speci�c context is highly malleable’ (Farrington et al., 2012, p. 24). In other words, if the context speci�city of perseverance is taken into consideration and the focus is on academic tenacity, there is reliable evidence that perseverance is a malleable disposition or skill. The �ipside of this approach is that it complicates the question of transfer to other, non-academic contexts. Since academic perseverance is more speci�c than conscientiousness (de�ned as a trait), the malleability of academic perseverance does not imply that it will necessarily transfer to other contexts. Evidence is strong to support that demonstrating persistence in one activity does not necessarily translate into a persistent behaviour in other contexts. Yet, there is a recent construct that has gained currency in the psychology literature for its more generic conceptualisation than task speci�city. The concept of grit, as de�ned by Duckworth and her colleagues, refers to a relatively stable characteristic or trait of displaying continuous application towards tasks or perseverance on tasks (Duckworth, 2016). Findings on grit are contrasted. On one hand, research indicates that there is a signi�cant association between grit and school grades (i.e. academic performance) (Duckworth & Seligman, 2005). On the other hand, ‘despite the intuitive appeal of this idea, there is little evidence that working directly on changing students’ grit or perseverance would be an effective lever for improving their academic performance’ (Farrington et al., 2012, p. 7). In other words, the correlational nature of evidence between grit and academic achievement does not imply that students would bene�t from interventions and practices speci�cally aimed at developing their grit (compared to other forms of interventions on cognitive skills such as problem- solving, for instance). In addition to being correlational in nature, the literature on grit or tenacity offers little clear implication for educational practice. Conclusion From the skill-by-skill overview, it is apparent that the different skills are similar to one another in some aspects but not in others. Overall, two key challenges evident in the literature on most skills are the issues of domain speci�city and transferability. Both are relevant questions for educators and policy makers to consider in order to decide the suitability of speci�c skills for inclusion in teaching and learning. Before addressing these two issues, it is important to summarise some key observations from the discussion above. The �rst point to be made is that de�ning the constructs considered to be the ‘key skills’ is not a simple task. Various academic �elds are involved and there is no process in place for organising the convergence of approaches or agreeing on the meaning of terms. In fact, it appears that there is signi�cant overlap between several of the skills included under the ‘21st century skills’ banner, and several skills de�ned separately tend to measure the same thing, at least in part. This is the case for locus of control and self-ef�cacy, for instance (Judge et al., 2002). This absence of mutual exclusivity at KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 28 Education: Future Frontiers | Analytical Report the construct level may be a �rst way of explaining the fact that authors have frequently noted that complex relationships exist between skills. For instance, Farrington et al. remark that ‘many non- cognitive factors are mutually reinforcing and […] relationships are often reciprocal’ (2012, p. 11). Beyond this cyclical causation mechanism, key skills such as critical thinking and creativity, or motivation and metacognition have been described as conceptually interrelated in complex ways (Lai & Viering, 2012, p. 29). The next point to be made concerns the quality and volume of evidence available for the study of key skills. Research into these skills is less advanced than research into traditional academic skills such as numeracy and literacy. In particular, ‘non-cognitive’ competencies, as de�ned by the US Committee on Deeper Learning (i.e. inter-personal and intra- personal competencies), have received far less attention than those labelled as cognitive ones. One interesting �nding to emerge is that, on the whole, the key skills listed in this chapter are reported to be malleable to some extent, with students able to develop levels of skill expertise. On the other hand, our understanding of the acquisition of the different skills remains limited (Binkley et al., 2012). Most of the cognitive and non-cognitive skills reviewed here seem to be especially malleable during childhood, which has important implications for early educational interventions in particular. At the same time, recent evidence also suggests that the skills not measured by cognitive tests are malleable until a later age than cognitive skills measured by IQ tests (Kautz et al., 2014), which can have implications for the later years of schooling as well. This is encouraging for the prospect of systematically teaching these skills beyond the early years or primary education only. There is a large body of evidence demonstrating correlations between various skills and grades, test scores and academic achievement. A few studies have even developed convincing arguments of causal relationships between some of the skills and future student success at school and in the workplace. In his synthesis of prior meta-analyses, Hattie (2008) examined, among many other factors, the degree of association existing between a range of key skills and academic achievement. Based on an analysis of effect sizes, Hattie reported that among different skills and dispositions, motivation, engagement in learning, self-ef�cacy (and related self-concepts), and persistence (interpretable as a form of perseverance) were among the more signi�cant correlates with traditional measures of academic achievement (Hattie, 2008, 2015). Other skills such as creativity show only a limited association with achievement. The effect sizes available for the skills covered in this chapter are presented in Figure 2-6. Effect sizes of 0.4 are considered average (Hattie, 2008). Scores between 0.2 and 0.6 could be considered as modest, below 0.2 as low and above 0.6 as high. Creativity, motivation and communication are at the lower end of the scale, while problem solving, cooperative learning and metacognition are higher. Unfortunately, there is no estimation of the interrelationships between skills and their combined effects on achievement. This is a limitation, given that the constructs are not always clearly mutually exclusive and do not necessarily work independently of each other. The epistemological vigilance needed when interpreting correlational studies is an important methodological principle for the broader literature on all skills. The existence of correlational evidence between a skill (or disposition) and success (school or work or life) does not provide evidence or proof that focusing on interventions to help improve such skills will lead to future success, and ‘it is not even true that intense correlations are more likely to represent cause than weak ones’ ( Gould, 1996, p. 272). Finally, the existence of strong correlations between speci�c skills and success does not provide a logical foundation for intervention. The fallacy with this form of reasoning becomes evident with the example of social factors associated with valuable outcomes: the fact that higher socioeconomic status, for instance, is consistently associated with enhanced educational outcomes has never been used as a justi�cation to design interventions aimed at changing students’ socioeconomic status. KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 29 Education: Future Frontiers | Analytical Report However, the results still have two important implications. First, a complex range of skills, dispositions and attitudes, is associated with—and hypothetically in�uences—student outcomes. Second, all the different skills and dispositions reviewed in this chapter do not appear to be equally associated with student outcomes. In addition, while the volume of research on the potential relationships between different constructs and academic achievement is large, there is little detailed and robust work discussing the relationships between key skills and career success and community life, for instance. ‘The available research evidence is limited and primarily correlational in nature; to date, only a few studies have demonstrated a causal relationship between one or more 21st century competencies and adult outcomes’ (Pellegrino & Hilton, 2012, p. 65). More research is needed, especially research going beyond cross-sectional designs and moving towards longitudinal analyses. The �ndings from the case-by-case analysis have signi�cant implications for the various frameworks on 21st century skills. Most of the frameworks propose that the skills they retain causally in�uence student outcomes, conceived in a broader sense than academic achievement. As our analysis of skills shows, the scienti�c literature is unclear about the extent to which many of these skills (or dispositions) directly impact on academic performance. At the same time, it should be acknowledged that traditional measures of student performance may be too narrow to capture some of the potential bene�ts that a focus on these skills might bring. The authors of all �ve frameworks presented above recognise the hypothetical nature of the causal mechanisms they include in their conceptual maps, and this is certainly one of their strengths. The gap between frameworks and research evidence suggests that more research is needed to understand the complex web of student skills and dispositions that causally in�uences student outcomes. Research would also bene�t from examining how the relationships existing between different skills shape academic performance, rather than focusing on speci�c skills taken individually. It is worth considering the question of domain speci�city, which has emerged as an issue throughout the literature on key skills. This issue is as theoretical as it is empirical, since the de�nition Source: Hattie, 2015 Figure 2-6 Estimated relationship of selected skills with academic achievement based on effect sizes KEY SKILLS FOR THE 21ST CENTURY: AN EVIDENCE-BASED REVIEW 30 Education: Future Frontiers | Analytical Report of a given skill largely determines its context speci�city. Perhaps a more useful question to ask is not whether a skill is generic or speci�c, but rather which conception of a given skill is the most useful for educators and students to focus on. If we re�ect on the usefulness of different conceptions of key skills, it is possible that these are best conceptualised as context-based ‘dimensions of expertise that are speci�c to—and intertwined with—knowledge within a particular domain of content and performance’ (Pellegrino & Hilton, 2012, p. 3) rather than as generic skills applicable to many tasks in a wide range of dissimilar contexts. The question of domain speci�city directly leads to the issue of transferability. To what extent can a demonstration of skill expertise in a given lesson or subject be evidence of the ability to use the same skill in other contexts? For example, to what extent can it be assumed that perseverance demonstrated by students in a task will be likely to transfer into perseverance in a different class of tasks? How different can these other contexts or tasks be until the transfer becomes much less likely to occur? These questions have proven challenging to answer for researchers, and more investigations are needed in this area. A range of factors in�uences the ability to transfer across domains, including the learning model (understanding versus memorising), learning time, motivation, the approach to transfer (active versus passive) and the context of learning (Bransford et al., 2000). Transfer of learning is more likely to occur between closely related domains (near-transfer) than between unrelated domains (far-transfer) (Mestre, 2005). Overall, there are signi�cant dif�culties in transferring knowledge and skills between school and non-school contexts (Sala & Gobet, 2016). Although certain approaches to teaching and learning can facilitate transfer between tasks or situations within a given context (or area), ‘over a century of research on transfer has yielded little evidence that teaching can develop general cognitive competencies that are transferable to any new discipline, problem, or context, in or out of school’ (Pellegrino & Hilton, 2012, p. 8). Perhaps the main lesson to be drawn from this statement is that there is a need to adopt a more reasonable conception of the transfer of skills students learn in classrooms. The transfer of general principles within subjects or curriculum areas has received encouraging evidence, at least when effective pedagogical practices are used. Therefore, so long as there is not held an overly ambitious conception of transfer, the accumulation of evidence suggests that knowledge and skills learned in school can transfer to related domains if attention is paid to the conditions of acquisition and transfer (Bransford et al., 2000). The conception of a skill as primarily context- dependent versus generic underpins the way skill transferability is approached, but it also raises the question of measurement. For researchers, the usefulness of a construct for understanding inter- individual differences is not entirely de�ned by the way this construct is measured. For educators, on the other hand, the instruments and techniques used for measuring student levels of expertise are crucial for the quality of educational practice. For any given skill to be considered as a legitimate learning outcome, teachers must be able to (1) systematically foster student skill acquisition in a learning progression, and (2) assess student levels of expertise for formative (and potentially summative) purposes. Skills can certainly be legitimate learning outcomes even if there is no way of measuring student levels of expertise with standardised tests , but they can hardly be legitimate learning outcomes if there is no way of monitoring student skill acquisition. Finally, there is much that is not yet known about the key skills needed for the 21st century. The broad range of dispositions and skills students need to succeed in school and beyond can be unpacked with different frameworks of skills, and many of the constructs included in such frameworks display complex interrelationships between skills. Education systems around the world have attempted to come to grips with this complexity, and the next chapter illustrates the ways in which skills have been taken up by different systems.