Abstract
The concept of intentional learning is well established in the educational community. Intentional learning implies that there is a recognized learning goal and that those involved are rational agents who can make choices with regard to achieving that goal. There are typically two different kinds of people involved in a learning situation—learners and those supporting learners (teachers, tutors, trainers, intelligent pedagogical agents, etc.). It is important to recognize that the goals of learners and those who are designated to support them may differ. Optimal learning occurs when such differences are minimized. A second well-established concept in education is that of engagement. Educational research has established time-on-task as a significant predictor of learning—the more time that a student spends on a learning task, the more likely that student is to master that task. More recently, this concept has been expanded to include initiating and sustaining engagement along with motivation and volition. According to the widely accepted constructivist epistemology, a person creates internal representations (a cognitive activity) to make sense of things that the person experiences, especially things that are new or puzzling in some way. In this paper, we review the logical landscape of the cognitive perspective of intentional learning and argue that it is incomplete without also taking into consideration a more holistic account of human learning that includes noncognitive aspects involved in human experience, somewhat akin to embodied cognition. Our review provides reminders of research-based principles that can inform technology-enhanced efforts to support intentional learning and promote engagement. We conclude with a framework for integrating technologies that are likely to engage learners taking into account a variety of learning goals, situations, and learners. This review is preliminary and intended to suggest areas worth investigating to elaborate a more complete picture of how best to promote and support intentional learning and active learner engagement using available technologies.
Introduction
A widely accepted principle from constructivist epistemology is that people naturally create internal representations to make sense of their experiences (Jonassen, 2013; Spector, 2012a). This natural meaning-making process goes on continuously with or without conscious effort or focused attention. On some occasions, however, that process can become very focused and concentrated, such as during a challenging problem-solving dilemma or a decision-making situation with a number of possible alternatives without an obvious best choice. Such experiences demand conscious effort and go beyond the naturally occurring meaning-making process just mentioned. While they still require a meaning-making activity (e.g. constructing a meaningful internal representation of the situation), they also imply a conscious intention to resolve the problem or reach an acceptable decision. This subset of learning experiences can be therefore considered intentional (Bereiter & Scardamalia, 1989; Scardalmalia & Bereiter, 2006; Sinatra & Pintrich, 2003). The more general process of creating internal representations to make sense of experience and the more specific process of intentionally focusing on a goal in the context of a particular situation are usually considered cognitive processes, and both are pertinent to nearly all learning and instructional situations.
However, Clark (2008) and others have argued that external factors in the learning situation and the cognitive processes brought to bear by an individual in creating internal representations to make sense of the situation and achieve an acceptable solution that do not provide a complete account of learning. Rather, to account for individual variations in learning, it is often necessary to take into account the individual’s bodily and emotional characteristics and how they influence an individual’s perception and subsequent thinking. Bodily characteristics and emotions may influence some cognitive processes. There are many types of emotions (e.g. anger, anxiety, pride, fear, frustration, guilt, enjoyment, and shame), and they can arise at any time for any number of reasons. It is difficult to predict emotions, especially within the context of a learning situation. Bodily characteristics include physical impairments and more persistent things such as attention deficit disorder, which also can influence learning but which are easier to identify and take into account. The point here is that such noncognitive factors can influence problem-solving and decision-making activities in a learning situation; an account that only considers traditional information processing factors (cognition) and aspects of the external situation (the learning environment) may not explain why learning does or does not occur as intended. A learner’s success or failure is not well or fully explained only in terms of cognition and the external learning environment.
Moreover, assuming that reasonably accurate depictions or renditions of an individual’s internal representations (results of a cognitive process) have been acquired, these may not match the solution or decision of that individual. What then accounts for such a mismatch? One could argue that the re-representations of internal processes were incomplete or inaccurate, which may well be true in some cases. However, in other cases, a mismatch could be accounted for by looking at noncognitive processes, including emotions or what might be called lower-level bodily factors. An example of the former would be a person who admits that a logical reaction would be X but the individual for whatever reason selects a dramatically different reaction, often based on an emotional factor (e.g. deep seated anxiety about some aspect of X and an anticipated emotion of regret for choosing X) (Schwarz, 1990). Emotions can be relevant to successful or unsuccessful learning, as many have argued (Kim, 2012; Kim, Park, & Cozart, in press; Kim & Pekrun, in press).
Bodily characteristics can also influence the learning and behavior of an individual immersed in a problem-solving or decision-making activity. For example, consider a task of taking apart a complex device, performing maintenance or repair, and then reassembling that device. In many cases, there exists a standard procedure for such an activity that might even be prescribed by the organization to ensure some level of quality control. However, a person with a missing finger or with a vision impairment might approach the task somewhat differently than the prescribed procedure indicates. Again, in order to account for the success or failure of the individual’s performance, more than external factors and cognitive processes might be required.
Learning is a complex human activity. The standard definition of learning is that it involves a stable and persistent change in what a person knows or is able to do (Gagné, 1985; Spector, 2012a, 2012b). While that definition is widely accepted, there is great variation in explaining how people learn (Bransford, Brown, & Cocking, 2000). A similar comment can be made about instruction. A standard definition of instruction is that it is an activity intended to support learning. While many will agree with that definition, there is a great deal of debate and discussion of how best to support learning in various situations. This variation certainly exists with regard to processes associated with intentional learning, learner engagement, and how best to support intentional learning and engagement. In the remainder of this paper, the focus is on learners and their active engagement in intentional learning situations. Our purpose is to provide a perspective and a framework that will guide productive research and hopefully will help those who wish to support intentional learning and promote active learner engagement in complex and challenging problem-solving and decision-making learning tasks. The cornerstone of our perspective and framework is a holistic view of the learner that includes both cognitive and noncognitive factors.
Research on intentional learning and learning engagement
Research on learning has established two different kinds of findings—namely, things that are widely accepted as influencing learning outcomes and things about learning that we still do not understand very well. There are many research findings that could be mentioned in either category. Only a few will be mentioned here prior to focusing on intentional learning and engagement.
Time on task is highly predictive of student learning and performance, according to a large body of empirical evidence (Brock, 2005; Ericsson, 2006). Practice should typically include feedback on performance and outcomes to keep learners focused on mastering desired outcomes. Timely and informative feedback is especially important for less-experienced learners. More-experienced learners require less feedback and become increasingly able to monitor their own learning (Collins, Brown, & Duguid, 1987). Moreover, time-on-task studies pertain to mastering psychomotor skills as well as cognitive skills (Ericsson, 2006).
A second well-established principle from research on learning is that prior student knowledge and experience influence the rate and quality of learning (Dochy, Moerkerke, & Segers, 1999). The notion that prior learning is significant is based on Anderson’s (1983) notion of how information is stored and retrieved in the mind. Basically, when something new is learned, it is connected to things already known and understood that seem relevant to the learner about the new knowledge. If a learner has little or no previous learned knowledge structures believed to be relevant, the newly learned information will be more difficult to retrieve and deploy in practice. Johnson-Laird (1983) developed the notion of a mental model to describe the process of what a learner does when confronted with a situation that demands the creation of a new internal representation or structure in order to respond to the situation.
This internal knowledge structure concept fits well with the notion of cognitive apprenticeship (Collins et al., 1987) as well as the notion of time-on-task. With regard to the latter, learning tasks that are practiced a great deal result in automated schema that are retrieved as entire knowledge chunks without conscious effort when needed. Many simpler procedural tasks that are performed more or less the same regardless of the surrounding conditions are now trained to the point of automaticity. However, such automatized knowledge, part of a cognitive structure, would not function automatically when noncognitive factors such as emotions interfere. For example, even after numerous practices of algebraic problems, a student’s mind could go blank during a test when experiencing anxiety.
Noncognitive factors can be used to get learners to spend more time on a learning task. Positive emotional experience can help learners become and remain interested in the task at hand. For example, the emotion of enjoyment enables flow, that is, full engagement, to occur (Csikszentmihalyi, 1988). Motivation (desire to learn) and volition (follow-through learning activity) are key to positive emotions and active engagement and thus to developing the relevant internal knowledge structures and mastering challenging learning tasks. While time on task is critical for learning, as research strongly suggests, the role of other factors, some of which are noncognitive, becomes quite relevant.
A cognitive perspective on learning
It would be difficult to argue that there is a single cognitive perspective on learning. However, it seems reasonable to construct a representation of a cognitive perspective based on common aspects of a variety of cognitive perspectives. One starting point is to begin with the nature of things to be learned (cf. Anderson, 1983; Gagné, 1985; Jonassen, 2013; Merrill, 2013). One could distill out the distinction between declarative knowledge (knowing that, as in facts and concepts), procedural knowledge (knowing how, as in the application of rules and principles in a particular situation), and contextual knowledge/causal understanding (knowing why, as in the ability to explain to others, identify assumptions and implications, and formulate alternative solution approaches). There are admittedly differences with regard to each of these categories. For example, different concepts appear subject to different interpretations (Frensch, 1998). The relationship of declarative and nondeclarative knowledge to both explicit knowledge and implicit knowledge should be considered but often is not (Kirkhart, 2001). The disaggregation of knowledge into discrete knowledge components does not necessarily imply that they should be taught separately.
Moreover, there are many who argue that what is to be learned does not always need to be separated into such discrete items. Gagné and Merrill (1990) argue that what is to be learned is typically a combination of related items that they called an enterprise. Van Merriënboer and Kester (2005) discuss the significance of learning whole tasks and using technology to support whole-task learning. Advocates of problem-based learning have long advocated introducing facts, concepts, rules, and principles in the context of an authentic problem that requires the application of the integrated collection of relevant information (Hmelo-Silver, 2004). Still, others argue that the collection of relevant information needed in a typical situation is often distributed among a collection of intelligent agents (Clark, 2008; Lave, 1988).
While these differences exist, one cornerstone of a cognitive perspective is that what is to be learned can be identified and represented in some form that is recognized by many as reasonable and legitimate. It is often the case that the desired representational form is computational in nature, especially by those cognitivists associated with an information-processing model of cognition (Anderson, 1983). The information-processing model posits the existence of externally recognizable knowledge components that can be presented to learners who then perceive, store, and recall these knowledge objects as they solve problems and make decisions. While such a model does not explain everything about learning, it does present a useful point of departure for those designing instruction as well as for those investigating learning (Sweller et al., 1998).
With regard to the latter group of cognitive scientists investigating learning, the prominent cognitive perspective that has emerged involves the notion of mental models—internally constructed representations created to explain or predict unusual or rare phenomena and situations. If one already understands all of the relevant factors perceived in a situation, then there is no need to construct such a representation or mental model. Mental models are just-in-need internal objects. They are entirely hypothetical in nature as no one ever directly observes a mental model—not even one’s own. However, it is widely accepted that the generation of productive mental models is crucial for problem-solving and higher-order reasoning. Such mental models are also suggested by a constructivist epistemology (Jonassen, 2013; Spector, 2012a). As with the information-processing model of cognition, mental models are useful to instructional designers in the form of devising the ways and means of promoting productive models as well as to researchers investigating the nature and development of expertise and problem-solving (Ericsson, 2006; Pirnay-Dummer & Ifenthaler, 2011; Scarmaldia & Bereiter, 2006; Schnotz & Kirschner, 2007).
Given these aspects of a cognitive perspective on learning and the substantial body of research conducted within that perspective in the last 50 or more years, one might wonder what is missing and what more needs to be investigated. Just as a behaviorist perspective fell short of explaining all aspects of learning (e.g. the acquisition of language by young children), a cognitive perspective also fails to explain much—for example, why well-trained and properly motivated adults with relevant information available often make terrible decisions (Dörner, 1996). There is a growing body of research that suggests that noncognitive factors influence both perception and cognition. These are discussed next.
Noncognitive learning factors
Problem-solving and decision-making processes are influenced not only by cognitive factors such as mental models but also by noncognitive factors such as emotions (Friedman, 2004). Because emotions can change the way information is stored and retrieved, information-processing, which is critical in solving problems and making decisions, can be altered (Levine & Pizarro, 2004; Linnenbrink, 2006; Schwarz, 1990). For example, it is likely that people recall their past experience that is congruent with their current emotions. This can be explained by the concept of emotion-congruent retrieval (Blaney, 1986; Bower, 1981; Levine & Pizarro, 2004; Parrott & Spackman, 2000). So, a student may recall study experience in a forensic psychology class positively (e.g. the emotion of enjoyment) while watching his favorite TV crime show. In some cases, taking into account noncognitive factors can enable a cognitive account to more completely represent an individual’s progress.
The impact of emotions can be crucial in intentional learning because it is related to goal-setting and goal-pursuit (Kim & Pekrun, 2014). For example, the goal to become a psychologist might be encouraged by building on memories from a prior forensic psychology class that a student found interesting. Emotions can also accelerate or interrupt the process of pursing a goal. The emotion of hopelessness can disengage students from complex and challenging problem-solving and decision-making learning tasks. However, negative emotions are not always against engagement and learning. They can be constructive; for example, shame can produce resiliency from challenging contexts (Turner & Schallert, 2001).
Understanding motivation is a key to fostering positive emotional experience. However, the influence between motivation and emotions is reciprocal (Kim & Pekrun, in press). For instance, Kim and Hodges’ (2012) study showed a strong positive relationship between enjoyment and motivation and a strong negative relationship between anxiety and motivation. And, in some learning contexts, the role of emotions is even more critical than the role of motivation; Kim et al. (in press) found that self-efficacy (i.e. a motivational factor) was not a predictor of student achievement in online math courses when emotions were entered in the analysis.
Motivation along with volition can be more powerful in leading to positive emotional experience and active engagement. Volition refers to the power that makes people keep striving to achieve their goals (Corno, 1993, 2004; Heckhausen, 2007; Keller, 2008; Kim & Bennekin, 2013). Some may argue that students already have intentions in the process of intentional learning and that should be sufficient to move toward intentional learning. However, intentions do not guarantee intended behaviors. Implementation of intentions is required to see an actual effect in the form of learning (Gollwitzer, 1999). Volition can also compensate for insufficient motivation (Kehr, 2004), which is observed when students encounter difficulties, especially during complex and challenging, problem-solving, and decision-making learning tasks. When students are not competent or confident in exploring alternative solutions for difficult problems, those with volition seek out help rather than abandoning the learning task. Emotions can be controlled with volition as well (Corno, 1993; Corno & Kanfer, 1993). Unwanted emotions such as feeling hopeless can be reduced using volitional control strategies (Kuhl, 1987, 2000). In summary, volition (including self-regulation and metacognitive skills) can be learned and used to direct, control, or manage an individual’s cognitive and noncognitive processes.
A framework for engaging intentional learning technologies
An instructional design framework generally consists of a number of principles drawn from the relevant research literature along with prescriptions for applying these principles to solve specific learning challenges or problems. An excellent example is the cognitive apprenticeship framework (Collins et al., 1987). The cognitive apprenticeship framework is based on such principles as: (a) students learn by doing, (b) advanced students require less support that those new to a domain, and (c) learning tasks should be perceived as meaningful and relevant to future activities. These principles are generally consistent with situated learning (Lave, 1988) and with Jonassen’s (2013) first principles of learning (Jonassen et al., 2008).
In addition, research on metacognition, self-regulatory learning, and volition is relevant (Corno, 1993, 2004; Efklides, 2006; Gollwitzer, 1999; Heckhausen, 2007; Heckhausen, & Beckmann, 1990; Keller, 2008; Kuhl, 1987; Pintrich, 2004). We focus on volition as an all-embracing construct that includes metacognition and self-regulation but also includes affective factors. We cannot offer a comprehensive review here due to the limited space but a simplified summary would consider volition “process-oriented visions” that focus on not only goal-setting but also goal-striving (Heckhausen, 2007; Kuhl, 1987, p. 279). A framework for intentional learning technologies that can be applicable here is Kim’s (2013) volitional control support design model that consists of (a) goal initiation (“Want it”), (b) goal formation (“Plan for it”), (c) action control (“Do it”), and (d) emotion control (“Finish it”). This model was constructed based on theories and research on volition and implemented in the design and development of a virtual change agent to improve students’ emotions and motivation in an intentional learning context. These four phases are embedded in the principles included in Table 1. Pintrich’s (2004) framework identifies four phases of self-regulated learning (forethought/planning/activation, monitoring, control, and reaction/reflection) and includes four areas for regulation in the framework: cognitive, affective, behavioral, and contextual factors. As Efklides (2006) argued, self-regulation is a complex and challenging task, which might require a learner to estimate the effort and time required to successfully complete a task—metacognitive skills. Developing metacognition, self-regulation, and volition are important in a holistic approach to learning as represented in our framework. Likewise, Pintrich (1999) argued for more research within a holistic perspective as follows: As noted by several of the articles, much of the research on self-regulated learning has been overly concerned with cognitive control and regulation, especially the different cognitive, metacognitive, and learning strategies that students may use to control their own cognition and learning. In contrast, these articles propose that there is a need for more theory and research on how students control their own motivation, emotions, behavior (including choice, effort, and persistence), and their environment. This renewed focus on the "whole" person and how an individual may attempt to regulate all aspects of the self and the context is a very welcome addition to research on academic self-regulation. (p. 336) Preliminary instructional framework for technology-enhanced intentional learning. Note: This framework is structured so that it can be extended to include additional principles, prescriptions, examples, technologies, and recommendations for research. It is intended to include both cognitive and noncognitive factors likely to influence learning and instruction. The first two columns are similar to previous frameworks and easily extended. The last three columns can be modified for specific learning situations and contexts. Beginning calculus was chosen for some of the elaboration since so much research has already been conducted with beginning calculus courses for non-math majors.
General prescriptions associated with these principles include: (a) teach both explicit and tacit content; (b) situate learning tasks in the context of meaningful and authentic problems; (c) demonstrate and explain processes; (d) provide feedback to students and coach their progress; (e) have students explain their work and their reasoning; (f) have students critically evaluate their work; (g) have students formulate hypotheses and explore alternative solutions; (h) present content in a simple to complex sequence; and (i) fade feedback and support as students gain competence and confidence. The prescriptions can be found in other instructional design frameworks and are generally consistent with Merrill’s (2002, 2013) first principles of instruction.
Model-facilitated learning (Milrad, Spector, & Davidsen, 2003; Spector, 2012b) extended the cognitive apprenticeship framework to the domain of complex and dynamic problem-solving domains. In doing so, the notion of graduated complexity was further elaborated to include not only sequencing from simple to complex in terms of content but also in terms of exercises and learning tasks while still retaining the concept of working on realistic problems as they might occur in a particular domain. For example, an air-traffic controller receiving initial training might be asked to direct a single aircraft in the landing pattern to land on a runway with no current activity and no problems. A more complex exercise is to direct two different kinds of aircraft in the landing pattern to land. Likewise, a programming student might first be asked to complete a particular sequence of code with a single missing element. Eventually, that student might be asked to develop an entire sequence of code in response to a similar problem. The notion of graduated complexity (Milrad et al., 2003) in the model-facilitated learning framework is consistent with the notion of situated learning and whole-task problem solving (van Merriënboer & Kester, 2005).
What is missing in these and other such frameworks are details pertinent to specific problem-solving and decision-making tasks, individual learner situations, and technologies likely to improve learning and instruction. The framework sketched in Table 1 is an initial step toward developing a more comprehensive framework that retains well-established instructional principles while adding more recent cognitive and noncognitive factors. The thread running through this framework is the notion that a comprehensive account of learning and instruction treats learners as humans with all of the complexity that implies. The perspective is intended to be holistic with regard to learners as well as with regard to learning support. Our framework is structured to include principles and general prescriptions as in the frameworks mentioned, but it also includes examples of more detailed prescriptions pertinent to specific learners and learning situations. In addition, relevant technologies and research studies that might be conducted are suggested. Of course, such a framework is not at all exhaustive. Rather, it is meant to be suggestive of how one might proceed in articulating the principles of situated cognition and prescriptions of cognitive apprenticeship and model-facilitated learning in the general context of technology-enhanced intentional learning environments.
Concluding remarks
The last column in Table 1 suggests a variety of research studies that are worth conducting to advance what is known about how technology can improve engagement in intentional learning situations. In many of these cases, mixed methods research that looks at quantifiable outcomes together with interview data and both teacher and student reports and observations is needed. There is less available research on the impact of noncognitive factors on learning and performance than the research that exists for cognitive factors, and there is still less research on the interactions that might exist among cognitive and noncognitive factors. While it seems likely that technologies will continue to involve and transform the nature of intentional learning environments, it seems equally likely that learner engagement over a sustained period of time will remain a critical factor in nearly all learning and instructional contexts and situations and especially in technology-enhanced intentional learning environments.