The Nature and Importance of Engagement

While implicit learning is recognized as constituting the major part of language acquisition, conscious mental involvement has also long been recognized as essential to efficient learning of novel material (N. C. Ellis, Reference Ellis and Rebuschat2015). Paying attention is important, whether one is trying to understand something unfamiliar or complex, problem-solve, or learn a different way of doing something (Baars, Reference Baars1997). Researchers of L2 acquisition have emphasized the need for L2 learners to pay attention to the connections between language form and its meanings in use (e.g., Gass, Reference Gass, Doughty and Long2003; Leow, Reference Leow2015; Long, Reference Long, Ritchie and Bhatia1996; Schmidt, Reference Schmidt and Robinson2001). This research has tended to focus on the construct of noticing, differentiating between noticing at the level of detection, awareness, and understanding (Leow, Reference Leow2015; Robinson, Reference Robinson1995). Previous research recognizes gradations of cognitive involvement, and teachers and researchers alike use the word “engagement” as a near synonym. However, as we will see here, paying attention is just one dimension of engagement.

Engagement refers to a state of heightened attention and involvement, in which participation is reflected not only in the cognitive dimension, but in social, behavioral, and affective dimensions as well. A seminal article on school engagement by Fredricks, Blumenfeld, and Paris (Reference Fredricks, Blumenfeld and Paris2004) described engagement as a “multifaceted” or “multidimensional” construct that includes, at the least, three components: cognitive, behavioral, and emotional. In applied linguistics, each of these and other dimensions are recognized as important to instructed language learning (e.g., regarding affect: Schumann, Reference Schumann1997; Swain, Reference Swain2013; regarding social factors, see Philp & Duchesne, Reference Philp, Duchesne, Philp, Oliver and Mackey2008), yet, each tends to be considered in isolation. Increasingly, researchers acknowledge the need to take account of the interdependence of these different facets of human experience (e.g., Larsen-Freeman & Cameron, Reference Larsen-Freeman and Cameron2008). In the education literature, these multiple dimensions are demonstrated to be overlapping and interdependent, not isolated independent constructs (Christenson et al., Reference Christenson, Reschly and Wylie2012). For example, when people are involved in a learning activity, experience is more memorable when affective states are also aroused (McGaugh, Reference McGaugh2013; Pekrun & Linnenbrink-Garcia, Reference Pekrun, Linnenbrink-Garcia, Christenson, Reschly and Wylie2012; Weiss, Reference Weiss2000). In contrast, the student who is bored or disinterested in a task is emotionally disengaged. Similarly, someone who is disconnected with other group members, and thus socially disengaged, may also be behaviorally off-task: not listening to responses of other members, not contributing to the interaction. They are unlikely to invest effort or persistence, or to direct attentional resources in effective ways to be cognitively engaged or even to fully complete the task (i.e., to be behaviorally engaged). Analysis of engagement allows us to include an emphasis both on attention (the cognitive dimension) and on the affective, behavioral, and social dimensions that support effective learning.

Christenson et al. (Reference Christenson, Reschly and Wylie2012) underscored the crucial role of engagement for learning: “Student engagement drives learning; it requires energy and effort; is affected by multiple contextual influences; and can be achieved for all learners.”(p. 817; see also Gettinger & Ball, Reference Gettinger, Ball, Thomas and Grimes2007). If we can understand engagement better, we are better equipped for investigating how to engage all learners. Engagement is a construct closely related to motivation. Indeed, it is described by some as the visible manifestation or “descriptor” of motivation (Ainley, Reference Ainley, Christenson, Reschly and Wylie2012; Cleary & Zimmerman, Reference Cleary, Zimmerman, Christenson, Reschly and Wylie2012; Martin, Reference Martin, Christenson, Reschly and Wylie2012; Reeve, Reference Reeve, Christenson, Reschly and Wylie2012; Schunk & Miller, Reference Schunk, Miller, Pajares and Urdan2002) and by others as the precursor of motivation (Pekrun & Linnenbrink-Garcia, Reference Pekrun, Linnenbrink-Garcia, Christenson, Reschly and Wylie2012). Anderman and Patrick (Reference Anderman, Patrick, Christenson, Reschly and Wylie2012) suggested that for this reason it needs to be explored through context-specific research, through observation of students in action, that is, in the classroom context, actually working on tasks in class. Detailed discussion of the relationship between motivation and engagement is outside the scope of this article. We note here, however, that it is a close and complex relationship (for further discussion, see Reschly & Christenson, Reference Reschly, Christenson, Christenson, Reschly and Wylie2012).

The Construct of Engagement

Descriptions of engagement tend to foreground characteristics such as interest, effort, concentration, active participation, and emotional responsiveness. That is, engaged students are not just going through the motions; they expend focused energy and attention, and they are emotionally involved. For example, Skinner and Pitzer (Reference Skinner, Pitzer, Christenson, Reschly and Wylie2012) characterized engagement as “constructive, enthusiastic, willing, emotionally positive and cognitively focused participation with learning activities in school” (p. 22). The interdependence of the dimensions of engagement is a vital characteristic of the construct. Before we discuss this interdependence in greater detail, we will first identify the salient characteristics of each dimension in turn.

Cognitive Engagement

Cognitive engagement involves processes such as sustained attention and mental effort (Helme & Clarke, Reference Helme and Clarke2001), often including self-regulation strategies. Helme and Clarke (Reference Helme and Clarke2001) identified a range of indicators of cognitive engagement in collaborative activities, including questioning; completing peer utterances; exchanging ideas; making evaluative comments; giving directions, explanations, or information; justifying an argument; and making gestures and facial expressions. Further indicators of cognitive engagement could include private speech and exploratory talk (see Barnes, Reference Barnes, Mercer and Hodgkinson2008; Mercer & Dawes, Reference Mercer, Dawes, Mercer and Hodgkinson2008). Evidence may come from audio and visual data, lesson transcripts, and observations, or indirectly through retrospective questionnaires and interviews such as stimulated recall (see Gass & Mackey, Reference Gass and Mackey2014), carried out postlesson.

An illustration of cognitive engagement identified via retrospective interviews comes from a study by Early and Marshall (Reference Early and Marshall2008) of a high school English class in Canada. These adolescent students, with English as an additional language (EAL), worked in groups to visually portray the meaning of a short story. The teachers and researchers found that having to personally express the key ideas of the story using multimodal devices enhanced the students’ level of involvement with the text, both behaviorally (through multiple readings and discussions) and cognitively (understanding, looking deeper). In the following excerpt from an interview, one student commented:

Cognitive engagement can be demonstrated by phrases such as “I think,” by causal connectives such as “because,” and by questions. There may also be evidence of argument or disagreement that reflects reasoning or exemplification. In Example 2, from Gibbons (Reference Gibbons1991), two primary school boys with EAL are involved in the task of designing a platypus enclosure for a zoo. Their cognitive engagement with this task is seen in the exploratory nature of their suggestions (Barnes & Todd, Reference Barnes and Todd1995; Mercer, Reference Mercer1995). It is also reflected in pauses, hesitations, reformulation, and repetition (lines 1–3). They build on one another's suggestions, each responding with new ideas and giving reasons (lines 4, 5, 6).

Whereas this example involves a focus on content, in Example 3, among adolescent foreign language learners of Spanish (Toth, Wagner, & Moranski, Reference Toth, Wagner and Moranski2013) the focus of engagement is on language itself as the students try to work out the underlying rules governing the use of inchoative se in a report of an earthquake and storm in Spanish.

Each of these examples indicates concentrated effort from the participants as they think about something they need to solve: how to use symbols, images, and words to synthesize a story (Example 1); what should go in a platypus enclosure (Example 2); and how to explain the underlying rule for use of a particular linguistic form (Example 3).

Of course, cognitive engagement is not only manifested in verbal form but also may be seen in facial expressions and body positioning, although verbalization of thought processes allows it to be more evident to the observer. Besides qualitative data sources, there are also quantitative research instruments that investigate engagement. Two examples of survey instruments created to explore engagement at the level of school use Likert scales and include items relevant to cognitive engagement at the class level. In the High School Survey of Student Engagement, (http://ceep.indiana.edu/hssse ), Yazzie-Mintz and McCormick (Reference Yazzie-Mintz, McCormick, Christenson, Reschly and Wylie2012, p. 750) included questions about U.S. students’ effort, investment in work, and learning strategies. Darr (Reference Darr, Christenson, Reschly and Wylie2012, p. 713), described a survey for New Zealand schools to investigate students’ self-perception of engagement. It includes items such as “I pay attention in class,” “I find it easy to concentrate on what I am doing in class,” and “I take notice of the comments my teacher makes about my work.”

Behavioral Engagement

Behavioral engagement is typically described simply in terms of time on task or participation. Thus being “on-task” is synonymous with behavioral engagement. Gettinger and Walter (Reference Gettinger, Walter, Christenson, Reschly and Wylie2012), based on research in U.S. schools, claimed that “academic engaged time,” that is, the amount of time students are actively involved, predicts academic achievement, and engagement is directly related to learning outcomes (see also Fredricks et al., Reference Fredricks, Blumenfeld and Paris2004). Similarly, in earlier SLA research, measures of engagement involving word counts (Bygate & Samuda, Reference Bygate, Samuda, Mackey and Polio2009) and turn counts (Dörnyei & Kormos, Reference Dörnyei and Kormos2000) are measures of behavioral engagement (time on task), without capturing other dimensions (e.g., cognitive, social, or emotional engagement).

In the following examples (unpublished data from Oliver, Philp, & Mackey, Reference Oliver, Philp, Mackey, Philp, Oliver and Mackey2008), two pairs of young children (ages 5–7) with EAL work on the same task to identify and count specific shapes in a picture. In Example 4, the pair is on-task behaviorally, focused on doing the task—they take turns identifying shapes in the picture and tally the numbers for their handout. In Example 5, Learner G is drifting off task and amusing himself with silly answers, to the frustration of the partner who remains on task.

Some researchers (Anderson, Reference Anderson1975) perceive behavioral engagement as a dichotomy: engaged (i.e., often synonymous with on-task behavior) or disengaged (i.e., associated with off-task behavior). Others (Finn & Zimmer, Reference Finn, Zimmer, Christenson, Reschly and Wylie2012; Mahatmya, Lohman, Matjasko, & Feldman Farb, Reference Mahatmya, Lohman, Matjasko, Feldman Farb, Christenson, Reschly and Wylie2012) consider behavioral engagement a continuum, depending on degree and quality of participation, using amount of effort, persistence, and active involvement as indicators of this. These indicators broaden the understanding of what constitutes behavioral engagement, and they help to provide theoretical explanations of links to learning and to other dimensions of engagement. Behavioral engagement has been measured qualitatively via observation of participation and effort as well as teacher reports and student self-reports or interviews (see Fredricks & McColskey, Reference Fredricks, McColskey, Christenson, Reschly and Wylie2012). In task-based studies, indicators tend to focus on academic behaviors such as answering questions or participating in tasks. Quantitative measures of task engagement may include items in survey instruments relating to participation and effort, for example, “I take care that my homework is done properly” (Darr, Reference Darr, Christenson, Reschly and Wylie2012, p. 713). Other examples of items include “When I am in class, I just act like I am working” (a reversed item) and, for persistence, “If I have trouble understanding a problem, I go over it again until I understand it” (Fredricks & McColskey, Reference Fredricks, McColskey, Christenson, Reschly and Wylie2012, p. 770). An example of an item on effort is, “I take care that my homework is done properly” (Darr, Reference Darr, Christenson, Reschly and Wylie2012, p. 713).

Emotional Engagement

The construct of emotional engagement is defined variously according to the research focus. For example, concerning school engagement, Yazzie-Mintz (Reference Yazzie-Mintz2009) described emotional engagement as “students’ feelings of connection to (or disconnection from) their school—how students feel about where they are in school, the ways and workings of the school, and the people within their school” (p. 16).

In relation to the context of the class and the task, Skinner, Kindermann, and Furrer (Reference Skinner, Kindermann and Furrer2009) defined emotional engagement as motivated involvement during learning activities, and they identified enthusiasm, interest, and enjoyment as key indicators of emotional engagement, and at the other end of the scale, anxiety, frustration, and boredom as indicators of negative emotional engagement (disaffection). Baralt, Gurzynski-Weiss, and Kim (Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016) added purposefulness and autonomy as aspects of emotional engagement. Emotional engagement may also include students’ feelings of connection or disconnection with their peers in the class and, particularly, their task interlocutors. For example, Early and Marshall (Reference Early and Marshall2008) noted the positive benefits of group work for the students in the class: “For the most part, working in groups also had a strong affective component and facilitated the students’ investment in the task” (p. 388). Students were eager to talk about their task, and did so outside of class time, enjoying the social benefits of having a common interest and purpose. As one student said:

This last quote reflects interdependence between the dimensions of engagement: The student spends extra time on the task, out of cognitive interest (finding an answer) and for the pleasure of sharing a conversation with friends. Here, social engagement during group work is supported through positive encouragement and active listening from peers. Conversely, learners may be disheartened by peer exclusion, resulting in negative emotions and lack of engagement. This is reflected in a study by Baralt et al. (Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016) among young adult learners of FL Spanish. As seen here, one student's frustration at exclusion by her partner in an online chat session resulted in deliberate disengagement:

We see from these examples that emotional engagement relates to motivation and refers to the affective nature of learners’ involvement. Emotions, whether positive (e.g., interest, enjoyment) or negative (e.g., boredom, frustration), influence effort and strategies for learning. The interdependent nature of dimensions is evident here: The emotional dimension links behavioral, social, and cognitive facets, as Pekrun and Linnenbrink-Garcia (Reference Pekrun, Linnenbrink-Garcia, Christenson, Reschly and Wylie2012) argued. They conceived of emotions as intrinsic to rather than as a facet of engagement, describing emotions as either activating or deactivating of engagement.

Social Engagement

While not included in all models of engagement, we believe that in the context of instructed language learning, the social dimension to interaction should be foregrounded as a dimension of engagement (see also Svalberg, Reference Svalberg2009). Social engagement is closely linked to emotional engagement, particularly among child and adolescent learners where affiliation is a powerful social goal (Philp & Duchesne, Reference Philp, Duchesne, Philp, Oliver and Mackey2008), at a period when peers provide a unique context for learning (Hartup, Reference Hartup1992). In recent research on interaction, particularly from a sociocultural perspective, there has been a new emphasis on collaboration between peers working on tasks together. Storch's (Reference Storch2002) model of patterns of interaction, based on Damon and Phelps (Reference Damon and Phelps1989), has been particularly influential here because it draws attention to mutuality in particular as impacting on success of task-based interaction between peers.

Following Storch (Reference Storch2002), a growing number of researchers (Moranski & Toth, Reference Moranski, Toth, Sato and Ballinger2016; Sato & Ballinger, Reference Sato and Ballinger2012; Storch, Reference Storch2008; Toth et al., Reference Toth, Wagner and Moranski2013) have suggested that learners are likely to be more effective in language learning when they are socially engaged: that is, when they listen to one another, draw from one another's expertise and ideas, and provide feedback to one another. Cognitive, social, and behavioral dimensions of engagement are evident in Example 6, from Moranski and Toth (Reference Moranski, Toth, Sato and Ballinger2016, p. 306). This pair of adolescent learners varies in knowledge of Spanish as a foreign language, but benefits from one another's expertise as they discuss form-meaning connections for se and me. The mutuality between them is evidenced by the reciprocity in their discussion: each building on the other's turn. In Example 6, RaquelFootnote ¹ asks many questions, which Diego has to think about and explain (lines 10–24, 36–39). In turn, Raquel argues the point on a form she identifies as reflexive (lines 13–21).

In Example 7 (Philp & Duchesne, Reference Philp, Duchesne, Philp, Oliver and Mackey2008, p. 96), from a classroom study of a kindergarten child (Yessara) using EAL with her peers, we see social engagement among three young children in a kindergarten class. While writing out identical lunch orders at school, they mimic one another both in speech and in action. Such reciprocity and mutuality reflects learners’ social engagement with one another.

The Importance of Context in Defining Engagement

One contributing factor in the multiple definitions of engagement has been the wide range of contexts to which it is applied. Janosz (Reference Janosz, Christenson, Reschly and Wylie2012) argued that understanding how engagement is expressed differently in various contexts is important to understanding its determinants and outcomes in those contexts. This perhaps underscores how the facets of engagement are conceived differently according to the underlying theoretical position, and the setting of the research. Reschly and Christenson (Reference Reschly, Christenson, Christenson, Reschly and Wylie2012) suggested that this is reflected inevitably by the “jingle” (same term used to describe different things) and the “jangle” (same construct described using different terms) of engagement research. In this article, we limit our context to task-based interaction in the language classroom, with variation expected also according to age and classroom setting (including mode, e.g., computer-mediated or face-to-face; individual, teacher-fronted, or peer interaction with the task; and instructional focus).

Focusing on task engagement, we argue that what counts in engagement in a language classroom will be specific to the learning goals and processes of language learning. For example, in recent second language acquisition (SLA) research on engagement, it is language-related episodes (Swain & Lapkin, Reference Swain, Lapkin, Bygate, Skehan and Swain2001) that are the predominant means of operationalizing engagement because of the importance placed on learners’ focus on language form, meaning, and use. Social engagement is another aspect that is particularly important to language learning, given the opportunities that social interaction offers for language practice. The particular processes by which this facet of engagement is linked to learning outcomes are different from those that link it to learning in mathematics, for example, where social interaction may provide opportunities to be exposed to other ways of thinking about a problem and to elaborate thinking (Webb, Reference Webb, Hmelo-Silver, Chinn, Chan and O'Donnell2013). The processes that are the focus of research will also affect the indicators of each of the dimensions. Research conducted in mainstream classrooms has already found different sets of indicators of cognitive engagement for different curriculum areas including mathematics (Helme & Clarke, Reference Helme and Clarke2001) and reading (Guthrie, Wigfield, & You, Reference Guthrie, Wigfield, You, Christenson, Reschly and Wylie2012). Storch (Reference Storch2008) has identified a set appropriate to the language learning context, which we discuss in the section on “Engagement in Task-Based Interaction.”

Considering Engagement as a Multidimensional Construct

The multidimensional and interdependent nature of components of engagement is seen, for example, in group work, where the participants may be so focused on the procedural aspects of the task (behavioral engagement) that they are not involved cognitively; they approach it in a surface manner, without really trying to understand it (O'Donnell, Reference O'Donnell, Alexander and Winne2006). Similarly, as seen in Example 5, excitement when working together in a task involving group work (social and emotional engagement) may interfere with or distract learners from cognitive and behavioral engagement. In some situations, the same dimension supports engagement in other dimensions. For example, we saw in the excerpt from Early and Marshall (Reference Early and Marshall2008) the power of social engagement in group work to awaken emotional, cognitive, and behavioral engagement. Table 1 provides examples of some of the ways in which each dimension can mediate other dimensions of engagement, either positively or negatively. The concept of “flow” (Csikszentmihalyi, Reference Csikszentmihalyi1990)—or the all-encompassing involvement of the individual to the exclusion of all else—suggests the ultimate in engagement, where all facets are positively involved. It is also possible for the dimensions to compete with or negatively influence one another, as is further explored below.

Table 1. Mediating Effects of Dimensions of Engagement

The multidimensional, interdependent character of engagement described above has a number of implications for research. First, it suggests that a focus on one dimension, while useful, provides only a partial picture: We need to measure more than one dimension, if we are to capture the full complexity of engagement. Second, as Janosz (Reference Janosz, Christenson, Reschly and Wylie2012) argued, we need to spell out the relationships between the different dimensions of engagement and explore how each mediates and influences the effects of the others, through theory and research. There is a range of positions taken in the engagement literature about these relationships, and SLA researchers will need to theorize and investigate them in their own context. For example, Lam, Wong, Yang, and Liu (Reference Lam, Wong, Yang, Liu, Christenson, Reschly and Wylie2012) argued that researchers should avoid overlapping of the engagement dimensions in their definitions, while Pekrun and Linnenbrink-Garcia (Reference Pekrun, Linnenbrink-Garcia, Christenson, Reschly and Wylie2012) described this as an aspect of the complex and multidimensional nature of the concept and described dimensions as inherently interdependent, for example, “cognitive-behavioral” or “social-behavioral.” What is the evidence from task-based L2 research? Our examples (given above and in Table 1) suggest that the second is a more accurate picture, but the particular relationships between the dimensions may vary across contexts. At the level of task, how engagement manifests and how the different dimensions interrelate will vary in each context for differing age groups of students, in varying curricular contexts, with different patterns of participant interaction, and for different kinds of tasks. Studying these relationships within a particular context, such as task-based language learning, is important to illuminate the processes involved.

In Table 1, we provide examples of how each dimension may mediate the effect of other dimensions. We now focus specifically on engagement in task-based interaction.

Engagement in Task-Based Interaction

Different facets of engagement are recognized in SLA literature, but to date, the construct itself is rarely conceptualized or operationalized, and exceptions tend to be limited to capturing a single dimension of engagement. As noted above, earlier work in SLA on engagement at the class level depended largely on gross measures of quantity of talk among learners (Bygate & Samuda, Reference Bygate, Samuda, Mackey and Polio2009; Dörnyei & Kormos, Reference Dörnyei and Kormos2000). However, this captured only aspects of behavioral engagement. With regard to engagement with language specifically, Storch (Reference Storch2008) identified instances in which peers talk about language use or provide correction. These language-related episodes (LREs) (Kowal & Swain, Reference Kowal and Swain1994) have been adopted as the primary unit of analysis for engagement in much of more recent work (e.g., Baralt et al., Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016; Dobao, Reference Dobao, Sato and Ballinger2016; García Mayo & Azkarai, Reference García Mayo, Azkarai, Sato and Ballinger2016; Svalberg, Reference Svalberg2012). Storch used LREs to examine the extent to which participants notice features of language while reconstructing a written text. Storch differentiated between “limited” and “elaborate” engagement in describing learners’ metatalk. The latter describes how learners “deliberated over the language items, sought and provided confirmation and explanations, and alternatives” (Storch, Reference Storch2008, p. 100). Thus, in Storch's model, engagement is limited to cognitive engagement with language.

However, consistent with an understanding of learning as involving social, affective, cognitive, and behavioral influences, we argue that it is essential to recognize engagement as more than unidimensional. Research drawing on a wide range of theorists, including Lewin, Vygotsky, Bandura, and Bronfenbrenner, has clearly demonstrated the interaction of the social, emotional, cognitive, and behavioral domains in learning and development (e.g., Adolph & Berger, Reference Adolph, Berger, Borstein and Lamb2011; Geeslin & Long, Reference Geeslin and Long2014; Lantolf, Reference Lantolf, Gass and Mackey2012; Larsen-Freeman, Reference Larsen-Freeman, Gass and Mackey2014; S. Mercer, Reference Mercer2011; Pekrun, Goetz, Titz, & Perry, Reference Pekrun, Goetz, Titz and Perry2002; Posner & Rothbart, Reference Posner and Rothbart2000). Just as in learning and development, then, in engagement as well, these domains interact with one another to shape human experience, as we have seen.

Svalberg (Reference Svalberg2009; see also Svalberg, Reference Svalberg2012) is one of the few applied linguists to date to recognize a complexity beyond cognitive or behavioral engagement alone. Svalberg's suggested model of “engagement with language” does not make reference to behavior, but it does incorporate the possibility of cognitive, social, and affective facets, noting that “in addition to its cognitive aspects [engagement with language] crucially involves a range of social and affective phenomena” (Svalberg, Reference Svalberg2009, p. 243). While she does not specify engagement as multidimensional (as did Fredricks et al., Reference Fredricks, Blumenfeld and Paris2004), Svalberg clearly acknowledges interdependence between social, cognitive, and affective states. Her identification of these multiple aspects of engagement is a critical step forward in understanding engagement in language learning contexts, because it recognizes the complexity of attention, learning, and development, as well as the many factors that mediate how students perceive, interpret, and engage with what happens in the classroom and beyond.

Some researchers recognize complexity by foregrounding agency as a central component, as seen in some educational research. For example, drawing on self-determination theory (e.g., Deci & Ryan, Reference Deci and Ryan1985), Reeve (Reference Reeve, Christenson, Reschly and Wylie2012) included agency as an interrelated facet of engagement, alongside behavioral, emotional, and cognitive facets (see also Michell, Reference Michell2012). Reeve described agentive engagement as the learner's “proactive, intentional and constructive contribution into the flow of the learning activity” (p. 151). For instance, Reeve recognized agency in the student's own input, participation, and suggestions—“enriching the learning activity” rather than “passively receiving” (p. 153). This work is useful with regard to task-based interaction in that indicators are tied to the active, observable contribution of the learner interacting with other learners (or with the teacher in whole-class interaction). Svalberg's (Reference Svalberg2009) description of engagement with language aligns with this description. Svalberg foregrounded the agency of the learner when she identified the learner as “interactive and initiating” (socially engaged); and/or as one who “pays focused attention and constructs their own knowledge” (cognitively engaged); and /or one who has “a positive, purposeful, willing and autonomous disposition towards the object (affectively engaged)” (p. 247).

Examples of Indicators of Engagement in Task-Based Interaction Research

In this section, we illustrate the different ways in which engagement, as a multidimensional construct, is identified with reference to two recent studies involving task interaction. Based on Svalberg's (Reference Svalberg2012) model, the work of Baralt et al. (Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016) qualitatively compared engagement of 40 adult learners in a North American university, learning Spanish as a foreign language during task-based interaction under one of two conditions: (a) in a face-to-face classroom setting or (b) through synchronous computer-mediated communication. The task involved story retelling, using a comic strip as the prompt, and differed in complexity according to whether information about the character's motives was provided (thought bubble included) or divined (empty speech bubble). Data comprised transcriptions, chat logs, and questionnaires. Consistent with Svalberg (Reference Svalberg2012), Baralt et al. described engagement as a cognitive and/or affective and/or social state, and identified these through a range of indicators developed from the data. Transcripts of group work interaction together with a posttask questionnaire were coded for evidence of cognitive, social, and affective engagement.

Baralt et al.’s (Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016) coding of the data was based primarily on instances of LREs. This is illustrated in Table 2. For cognitive engagement, they identified noticing and reflection on language form. Affective engagement related to learners’ attitudes, willingness to interact, their purposefulness (orientation), and their autonomy (pp. 222, 227). This differs from the conception of emotional engagement in the education literature. Social engagement related to aspects of mutuality and reciprocity, which Baralt et al. described in terms of interactiveness and supportiveness, as well as relational indicators of friendship, trust, and inclusion (p. 229).

Table 2. Coding Scheme (adapted from Baralt et al., Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016, pp. 222)

Perhaps symptomatic of the interdependence between dimensions of engagement, categories often overlapped. For example, learner scaffolding (or lack of support thereof) was coded as social engagement—yet this could also constitute an LRE, an indicator of cognitive engagement.

The self-report questionnaire helped to illuminate learners’ perceptions of the interaction, and reflected varied engagement. Eight questions elicited statements about overall perception of the task, goal of the task, noticing of any specific language features, opinion as to usefulness of working with the partner, helpfulness of the partner, and the students’ relative willingness to contribute as a pair. It also elicited adjectives to describe the task and their feelings about the task.

In Example 8, one student's reflection provides evidence of both cognitive ^(COG) and social engagement ^(SOC) (Baralt et al., Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016, pp. 227, coding added).

Other responses reflect the emotional dimension of engagement, as seen in Example 9 (Baralt et al., Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016, unpublished data, and p. 228, respectively, coding added). One student's experience was framed very positively, suggesting a willingness to participate, positive attitudes and motivation; and for another it was negative.

The social dimension of engagement highlights social relations between learners as an important variable in task outcomes, something that often remains hidden in studies of task-based interaction because it is not considered.

Another study of task-based interaction compared engagement in dyadic narrative and opinion gap oral tasks, among 32 adult learners of English in Japan. For logistical reasons, Lambert and Philp (Reference Lambert and Philp2015) identified engagement based only on audio recordings and transcriptions of their pair work. Based on educational research (Christenson et al., Reference Christenson, Reschly and Wylie2012), Lambert and Philp described engagement as a multidimensional construct involving cognitive, behavioral, and social-emotional facets. Like Baralt et al. (Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016), indicators of engagement were developed from the interaction data, resulting in some variation according to task type. For example, in the narrative task, the participants’ responsiveness and attentive listening are reflected both through questions and negotiation of meaning (cognitive engagement) and by back channeling, commentary, and expressions of empathy (social/emotional engagement). In the opinion task, indicators of engagement included provision of reasoning, offering opinions, or acknowledging the interlocutor's opinion. Here “opting out” (e.g., by giving no response or saying “I don't know”) was an indication of low or no engagement.

In the narrative task, one participant was instructed to recount an interesting story that had happened to him or her personally. Although ostensibly a monologue, in most cases this was carried out dialogically, reflecting the engagement of both participants. In Example 10, the storyteller (S) talks about problems at the airport in Thailand when returning from Myanmar. Both questions and negotiation of meaning indicate cognitive engagement in this example, as the two participants work to understand one another (lines 1–8). In the opinion gap task, indicators of cognitive engagement included questions, reasoning, and suggestions, as well as negotiation of meaning. As seen in lines 9–11 of Example 10, social engagement is reflected in an extended closure to the story, and by empathy of the interlocutor (line 15), as she imagines the potential problems that might have ensued.

Example 11 involves a different pair of students performing the same narrative task as Example 10. It further illustrates coding for indicators of social and emotional engagement. In this case, the story, told by a female first-year student, involves illness related to stress and homesickness. Her interlocutor (FI), an older female student, empathizes with the speaker's problem, saying “I know how you feel” and “I was just like you,” and provides advice and encouragement.

Implications for Measures of Engagement in Task-Based Interaction

Understanding engagement as multidimensional suggests that our measures should not just reflect learners’ cognitive responsiveness but also social, affective, and behavioral aspects. In order to capture this, we are likely to require a range of indicators and sources, and these will differ according to context. In the specific examples of engagement in task-based interaction provided above, the main source was the transcripts of interaction, with some support provided through original sound files and exit questionnaires. Ideally, other sources would be used to complement these data. For example, Michell's (Reference Michell2012) study of learner engagement in activities in primary and secondary classrooms involved EAL learners, and it exemplifies how analysis of video recordings of classroom interaction can produce a wider range of indicators of engagement than transcripts alone. These include both verbal and nonverbal behaviors such as students’ gestures, speech, and action within a lesson. Michell identified animated talk, laughing, exclamation, mounting excitement, collective affect bursts, waves of excitement, raised volume, bilingual comments, self-commentary, and nonverbal expressions as “affect displays” (pp. 412–414), which we would describe as indicators of emotional engagement. He was also able to identify indicators of engagement in students’ “embodied dispositions,” such as leaning forward and moving closer to the focus of the task (Michell, Reference Michell2012, pp. 412–414). Data-driven measures of engagement such as this (as opposed to a list of predetermined indicators) are likely to provide the necessary flexibility to shape operationalization of engagement to the particular context of the research study in task-based research in classrooms. The use of motivation questionnaires (e.g., Kormos & Csizér, Reference Kormos and Csizér2014) and introspective interviews, such as stimulated recall (e.g., Baralt et al., Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016), would also provide insights into learner perceptions. Use of time sampling (e.g., Park, Holloway, Arendtsz, Bempechat, & Li, Reference Park, Holloway, Arendtsz, Bempechat and Li2012) may best capture fluctuations in engagement.

Each of the various methods of assessing engagement has strengths and limitations, which may make one method more useful for assessing one dimension rather than another. For example, Skinner et al. (Reference Skinner, Kindermann and Furrer2009) suggested that self-report or interview may be better than observation for assessing cognitive or emotional engagement, whereas observation may be better for assessing behavioral and social engagement. Some measures may also be more suitable for particular groups of participants—self-report surveys may have limited reliability with younger students, for example. Following Fredricks and McColskey's (Reference Fredricks, McColskey, Christenson, Reschly and Wylie2012) review of methodologies, in which they discussed the advantages and disadvantages of various measures, the use of multiple methods is recommended, in order to gain a full understanding.