Commentary: Bridging student engagement research and practice

My charge from the guest editors was to highlight the practice and policy implications that accompanied the five empirical articles that were featured in the April 2017 special issue on student engagement (Liem & Chong, 2017). As I proceeded with my appreciative review and critique of these articles, needs for a second purpose became apparent: To provide an integrative framework that connects these articles, identifies their inherent strengths and limitations, and describes other emergent research, practice, and policy priorities.

Background and conceptual rationale for the commentary

The April 2017 issue of School Psychology International (SPI) was organized to help school psychologists, educational leaders, and policy makers identify international best practices for fostering student engagement in mainstream classroom settings. This topic is timely and important. It reflects growing recognition that today’s international policy agenda, with its focus on college and career readiness for all, depends on the active, sustained engagement of all students (Christenson, Reschly, & Wiley, 2012). It also reflects the popular view that wholesale enhancements in student engagement may not be possible if educators and other helping professionals lack access to theoretically-sound, and research-supported educational practices, policies, and programs (Fuchs, Fuchs, & Compton, 2012).

I highlight the terms ‘theoretically-sound’ and ‘research-supported’ in the previous sentence because I think they aptly characterize much of the current relationship between engagement research and theory, on one hand, and educational practice and policy, on the other. In this current research-practice relationship, research provides practitioners with empirically ‘tested’ and ‘verified’ theoretical knowledge. This theoretical knowledge allows practitioners (and other scholars) to understand how school and classroom learning ‘works’ across an average set of school, family, and community conditions. Practitioners, in turn, are expected to use this theoretical knowledge much in the same way that academics do (Lawson, 1990). This means taking the time – whether it is personal time or time spent through sponsored professional development activities – to study the literature and learn its specialized terms and language. Once this specialized language is ‘learned’, the assumption is that practitioners will translate the theories and findings gained from research into practices that fit the needs of local students, classrooms, and schools (Schön, 1987).

Readers should know that I practiced and benefited from this kind of ‘technology transfer’ model during my decade-long career as a school-community practitioner (e.g., Lawson & Alameda-Lawson, 2012). And because of that work, I continue to believe that there is nothing as practical as a good theory (Weiss, 1995)! On the other hand, theoretical knowledge is not fool proof, and it doesn’t always transfer well to real world practice environments (Schön, 1987). In fact, in some cases, an over-reliance on theoretical knowledge can silence the practical needs and challenges posed by today’s educational practice and policy agenda (Bryk, Gomez, Grunow, & LeMahieu 2015).

Take today’s Response-To-Intervention (RTI) framework as a case in point. In this tiered intervention approach, the charge of school psychologists, educators, and other pupil support personnel is to enhance student access to a tailored system of learning supports (Eagle, Dowd-Eagle, & Holtzman, 2015). As a part of providing these services, school psychologists and other school professionals must assess and manage the engagement-related needs of particular students, at particular developmental ages and stages, in particular schools, in particular academic subjects and disciplines, given a particular set of school-community resources and constraints.

I highlight the word ‘particular’ here because I want to emphasize how this focus on providing individualized care often contrasts with the stated aims of much engagement research, including the studies featured in the SPI special issue. Like most extant engagement research efforts, these SPI studies were not designed to direct practice with individuals (or systems), nor were they meant to bestow professionals with the kind of tangible ‘know how’ that might guide their daily work. Instead, these studies were designed principally to enhance extant theoretical knowledge of how engagement relates to motivation and other features of instructional context.

To be clear: In framing this argument, I am not suggesting that a broad focus on describing, explaining, and predicting how engagement works generally is undesirable. To the contrary, theory testing and development is vital, not only for scientific knowledge (Kuhn, 2012), but also for professional learning and development (Fuchs et al., 2012). My point is that bridging connections between student engagement research and practice may require a new line of studies that take stock of the nuanced demands of practice in diverse school contexts – local, state/provincial and national. The articles in the SPI special issue provide important clues about what a companion, research-practice partnership agenda might look like. I return to this claim later in the commentary.

The downside of theoretical/disciplinary fragmentation

A second practice-related limitation posed by much of today’s engagement research lies in the fact that there are multiple, even competing engagement theories, and most are highly specialized. While all such specialization is an indicator of engagement research’s status as a vibrant, scientific enterprise, it also poses considerable practical challenges for practitioners who have little time or even the inclination to read academic journals.

For instance, a quick cursory review of the SPI engagement articles indicates that readers may need to be conversant in several different lines of motivational theory just to understand the abstracts. Then, in order to compare and contrast the articles and their findings, they need to unpack how and why researchers use different research terms to describe similar things (e.g., competence and self-efficacy), and they also need to reconcile how and why researchers use identical terms to describe qualitatively different things (e.g., why one article defines affective engagement as ‘academic interest’, while another defines it as ‘anxiety’). All in all, this theoretical diversity and complexity, an important strength for engagement research and scholarship, threatens to splinter the knowledge base for practice and policy at a time when it needs to be synthesized and integrated (Boekaerts, 2016).

To this point, Eccles (2016) suggested that one of the primary challenges facing engagement research is that too much research attention has been placed on studying particular engagement theories and/or sub-components instead of exploring the larger conceptual space which she calls engagement. This concern led her to question whether engagement researchers have lost sight of the forest as they have focused on particular trees.

My commentary follows suit. I too think that it is timely to take a more integrative, bird’s-eye view of the field. In service of this important agenda, I offer a high altitude, inclusive perspective of each article’s relationship to three vital types of engagement research theory and research. I offer this typology in service of two purposes: (1) to assist readers’ ability to categorize, classify, synthesize, and integrate engagement research, theory, concepts, and definitions; (2) to highlight the practice implications that follow each study. I conclude with a few recommendations for how future research might better attend to the day-to-day practice realities of school psychologists, educators, and other pupil support personnel.

Issues of conception and definition

According to Eccles’ (2016), much of the current research and practice interest in engagement was stimulated by a group of scholars who were convened by the MacArthur Foundation to better understand the psychological and social factors that influence children’s academic achievement during elementary school. Obviously, much has changed since that time. Today’s engagement researchers also tackle complex problems such as school failure, disengagement, and early school leaving (aka dropout). What is more, some researchers have adopted a life course developmental perspective on engagement with particular attention to how it might vary as a function of school level (e.g., elementary school versus high school).

Within this broad focus, most researchers conceptualize engagement as a meta-construct, consisting of affective-emotional, cognitive, and behavioral subcomponents (Fredricks, Filsecker, & Lawson, 2016). Although there is significant variation in how these terms are defined, emotional engagement typically refers to student feelings of identification and belonging to school, as well as the level of interest, enjoyment, happiness, boredom, and/or anxiety they experience conducting academic work (e.g., Pekrun & Linnenbrink-Garcia, 2012).

Cognitive engagement refers to students’ psychological investments in learning tasks (Fredericks, Blumenfeld, & Paris, 2004), the cognitive effort they exert while studying (Finn & Zimmer, 2012), and the extent to which they think deeply about academic ideas and concepts (Cleary & Zimmerman, 2012). Csikszentmihalyi (1999) describes the confluence of deep emotional and cognitive engagement as indicative of ‘flow experiences’. When students experience states of flow, they become so intently engaged cognitively and emotionally, they lose awareness of time and space.

Research on students’ behavioral engagement typically examines student participation in school and academic work (Fredericks et al., 2004). Some of this research focuses on indicators of students’ pro-social conduct in academic settings. Examples of these indicators include the amount of time students spend on homework and/or the extent to which they comply with school rules (Christenson et al., 2012). Other studies, particularly those in the drop-out literature, employ measures of student conduct as proxies for student disengagement and disaffection (Skinner & Pitzer, 2012)

Toward more precise and nuanced definitional frameworks

Although research has generally coalesced around the above-described, three-component view of student engagement, some researchers have continued to add nuance to extant definitions and concepts. These more recent definitions are important because they help researchers better attend to the quality of students’ engagement and attention, as well as the instructional context in which engagement occurs.

For example, prior research on students’ behavioral engagement has often been conducted with a conformity/compliance-oriented lens (Lawson & Lawson, 2013; Shernoff, 2013). This perspective is consistent with industrial age classrooms, where student engagement is viewed as passive and needing to be stimulated by a classroom teacher (Zhou & Ren, 2017).

In contrast to this compliance-oriented view, Reeve (2012) and others (e.g., Crick, 2012) have championed an alternative set of engagement-relevant concepts. These alternative concepts are relevant to classroom settings where teachers encourage students to actively express their thoughts, opinions, and interests with other students in the classroom. Emphasizing this more active view and reflexive view of engagement, Reeve coined and validated the social-psychological construct of agentic engagement (see also, Jang, Kim, & Reeve, 2016).

Other engagement definitions

Another recent development in engagement research involves the psychological construct of valence (e.g. Pekrun & Linnenbrick-Garcia, 2012). The notion of valence is important because it helps researchers distinguish between educational events that have intrinsic appeal to students (positive valence) versus those that yield adverse experiences and/or consequences (negative valence).

Recent research on student disengagement provides a good example of these conceptual developments. Specifically, whereas prior research considered engagement and disengagement as two sides of the same conceptual coin (i.e., they represent opposite ends of the same conceptual continuum), researchers are now treating these variables as independent concepts in their research designs (Jang et al., 2016). For engagement practice, this kind of differentiation has intuitive appeal. It suggests that that the practice strategies needed to promote engagement may be quite different from those needed to reduce disengagement (Lawson & Lawson, 2013).

Operational considerations and developments

The previous section highlights the multiple definitions of student engagement that are currently ‘at work’ in today’s research conversation (see also Reschly & Christenson, 2012). This diversity is important for practice because it sensitizes school psychologists, educators, and other school professionals to the variety and complexity of students’ everyday school experiences (e.g., Martin et al., 2015). However, the range of engagement-relevant concepts highlighted in the SPI special issue (and engagement research overall), makes it difficult to draw solid practice-related conclusions from extant research.

One way that readers can manage this theoretical complexity is to situate each SPI engagement article within particular lines or ‘types’ of engagement research and theory. Readers can identify these lines of research by using two analytic lenses. The first such lens focuses on the ways in which researchers ‘operationalize’ engagement – that is, how they characterize engagement’s role in the learning process.

Several operational views of engagement are present in the SPI engagement articles. These views depict engagement as: (a) a state of experience that facilitates student learning (Shernoff, Ruzek, & Sinha, 2017); (b) a behavioral variable that is influenced by certain kinds and degrees of student motivation (e.g., Martin, Martin, & Evans, 2017); and (c) an integrative, trait-based construct that subsumes motivation and other engagement-relevant concepts (e.g., Watt, Carmichael, & Callingham, 2017).

A second way the SPI articles can be analysed is to identify their relationship to particular psychological theories and traditions. Framed in this way, the five SPI engagement articles can be appreciated for their important contribution to five identifiable lines of engagement and motivational theory. These articles extend these theories to new student and school populations (including international ones); and, in so doing, help fortify three important types of engagement research.

Type 1 engagement research: Engagement as a state-based construct

Type 1 engagement research is anchored in positive psychology. It revolves principally around the psychological construct of flow. In these studies, engagement is operationalized as a ‘state’ of experience, meaning that it occurs at particular moments in time. Because these studies frame engagement as a highly variable and moment-specific phenomena (e.g., Martin et al., 2015), some scholars refer to it as situational engagement (Shernoff, 2013).

Framed in this way, most studies of students’ situational engagement are designed to capture students’ experiences while they are participating in particular educational activities. In these studies, students’ ‘in the moment’ experiences are typically measured by researchers through an innovative research strategy called Experience Sampling Methods (ESM). In this ESM approach, students are cued to record their engagement experiences during particular intervals of a classroom lesson. Once recorded, students’ experiences are analysed by researchers in ways that allow them to understand how engagement is influenced by particular instructional strategies or techniques.

The Shernoff et al. (2017) study represents an important example of this Type 1 engagement research. In this study, the authors used ESM to examine the situational engagement of 104 high school students in the United States. The goal of this research was to explore whether engagement might act as a ‘mediator’ between instructional context (which they refer to in the study as environmental complexity) and student’s perceived learning. The term ‘mediator’ is used in this study to highlight the research proposition that instruction will not impact student learning unless students are engaged (i.e., context→engagement→learning).

To examine engagement’s role in ‘mediating’ the instruction-learning relationship, the authors videotaped and then coded teacher practices for the presence of two novel instructional concepts (after Shernoff et al., 2016). The first concept, which the authors call environmental complexity, was indicated by teacher attention to students’ conceptual/language development, the presence of authentic and challenging tasks, and as other instructional strategies. The second concept, which they call environmental support, was indicated by teacher efforts to provide motivational support for learning, positive relationships, performance feedback, and so forth. Hierarchical Linear Modeling was used to analyse relations between these novel constructs, engagement, and students’ perceived learning.

Consistent with extant engagement theory (e.g., Christenson et al., 2012), results from this study support the widely held proposition that engagement can as an important ‘throughput’ (or mediator) between instructional context and students’ self-reported learning. What is more, the authors’ analysis of their environmental support variables appear to offer fresh insight into the instructional strategies that might help foster student engagement. Specifically, these results indicate that student engagement can be enhanced when teachers foster caring relationships with their students and when they promote positive motivational climate in the classroom.

Notwithstanding the practical merits of this important finding, Shernoff and colleagues’ (2017) study included a second set of results that call into question the utility of these findings for all classrooms. Specifically, the researchers found that when classroom-level differences were controlled in their statistical models, the relationship between engagement and environmental support became statistically non-significant. This null finding is important because it provides wholesale cautions against claims that there are singular ‘best practices’ for fostering engagement for all students, in all classrooms. In fact, findings from this study appear to support the notion that, where engagement, learning and instruction are concerned, one size does not fit all.

Type 2 engagement research: Engagement as a malleable (but increasingly durable) recursive process

While the Shernoff et al. (2017) article conceptualized engagement as a particular ‘state of experience,’ three other articles in the SPI special issue offer a different operational view. In these ‘Type 2’ engagement research studies, engagement is considered as a (mostly) behavioral variable that can help researchers explain complex relations between instructional context, motivation, and children’s academic achievement (Skinner & Pitzer, 2012). In brief, this Type 2 engagement research follows a process-oriented approach, which is grounded in particular theoretical traditions.

Self-determination theory

Self-determination theory (SDT) informs one line of process-oriented engagement research. SDT is based on the proposition that individuals have basic psychological needs for autonomy, competence, and relatedness (after Connell & Wellborn, 1991). When these needs are met, the motivational conditions for engagement are realized, and powerful, deep learning may be facilitated as a consequence. When they are not, engagement may be sub-optimal and, in the worst cases, disengagement may follow.

Overall, SDT’s attention to ‘getting the conditions right’ for motivation highlights the critical role that educators can play in creating classroom environments that are autonomy supportive. Among other co-requisites, these autonomy support environments can be witnessed when educators foster positive relationships in the classroom, and when they allow students to co-direct their own learning (e.g., Reeve, 2012). Following this view, the proposition that guides much SDT engagement research is that effective instruction will enhance student motivation, motivation will enhance engagement, and engagement will enhance learning and achievement. In research terms, this theoretical proposition can be expressed as context→motivation/basic psychological needs→engagement→achievement.

Zhou and Ren’s (2017) study of task disengagement in China extends this primary line of SDT research and theory in important ways. Drawing from recent research on the valence of student engagement (e.g., Jang et al., 2016), these authors defined students’ task disengagement as a behavioral variable that is conceptually distinct from students’ task engagement. The goal of their study was to explore the motivational and contextual-correlates of task disengagement among 347 fifth grade students.

To explore these relationships, the authors analysed several structural equation models. Each model explored task disengagement’s relationship with two different types of student motivation (i.e., autonomous motivation and controlled motivation) as well as two different kinds of learning orientation (competitive learning orientation – i.e., wanting to out-perform others, and collaborative learning orientation – wanting to work with others). Their final models analysed these constructs in the following way: Task disengagement was positioned as their primary outcome variable, and students’ collaborative and competitive learning orientations were specified as mediators of the relationship between student motivation and engagement.

This study revealed an important, practice-relevant causal chain of events. This chain begins with Zhou and Ren’s (2017) research finding that increases in students’ autonomous motivation can enhance student preferences to work with others. It then extends to their finding that related students’ collaborative learning orientations to reduced levels of task disengagement in school.

These results are important for practice because they highlight the potential role of collaborative learning in reducing students’ task disengagement in the classroom. To this point, prior engagement research has typically portrayed student engagement as a mostly individualistic process and phenomenon. However, results from this study indicate that engagement might be better understood by practitioners as a collective process that has socializing power (Lawson & Lawson, 2013). For practice, the immediate implication is that students’ task disengagement might be significantly alleviated through group work that targets students’ agentic engagement (Reeve, 2012). In an RTI framework, this means structuring Tier 2 intervention supports in ways that encourage students to actively express their thoughts, opinions, and interests with other students in the classroom (e.g., Jang et al., 2016).

Expectancy-value theory (EVT)

The second kind of Type 2, process-oriented engagement research originates from Jacquelynne Eccles’ work on Expectancy-Value Theory (e.g., Eccles et al., 1993). In EVT, student’s academic achievement and engagement are driven by two particular sources of student motivation: (a) student expectancies for academic success; and (b) their subjective task values. In this framework, student expectancies refer to self-perceptions of their ability and/or confidence to complete a particular task. Students’ subjective task values consist of four, inter-related component parts: (1) attainment value refers to the personal importance individuals place on particular academic subjects or tasks; (2) intrinsic value refers to the enjoyment or interest individuals assign to task engagement; (3) utility value refers to the perceived usefulness or relevance of task; and (4) cost refers to a task’s potential ‘drain’ on personal and psychological resources, including time.

Because research indicates that students’ school-related task values weaken during late childhood and throughout adolescence (Eccles & Wang, 2012), a key priority for EVT scholars has been to create learning environments that support students’ developmental strengths and needs. The concept of stage-environment fit characterizes learning environments that best accommodate student’s diverse developmental strengths, needs, and interests (e.g., Eccles et al., 1993).

The SPI study by Martin et al. (2017) examines engagement by way of one such EVT framework. The purpose of their study was to analyse associations between context, motivation, engagement, and academic achievement among a high school sample of 585 students in Jamaica. In this study, context was defined as students’ motivational milieu (i.e., students’ perceptions of the academic expectancies and values held by their parents, teachers, and peers); motivation was indicated by students’ school-related expectancies and task-values; engagement was defined as students’ pro-social school behaviors (i.e., participating in class; completing homework, and attendance); and academic achievement was measured by student test scores across several subjects. Structural equation modeling was used to analyse these constructs according to the following theoretical progression, i.e., context→motivation→engagement→academic achievement.

Consistent with extant engagement research and theory (e.g., Christenson et al., 2012), this study yielded significant findings along all paths analysed in their statistical models. For instance, students’ perceived motivational milieu was significantly related to their task values and expectations (i.e., their motivation). Students’ task expectancies were positively related to students’ behavioral engagement (homework completion and class participation) and negatively related to their behavioral disengagement (i.e., absenteeism). What is more, students’ homework completion and class participation were positively correlated with academic achievement, while student disengagement/absenteeism was negatively associated with this same outcome.

Two primary practice implications can be derived the Martin et al. (2017) study. The first implication follows their rather interesting finding that students’ task expectancies matter more for behavioral engagement than their subjective task values. This finding is important for engagement practice because it suggests that student perceptions of their academic efficacy (Shunk & Mullen, 2012) and competence (Reeve, 2012) represent primary drivers of their engagement and achievement. Moreover, because students’ prior achievement patterns have been shown to influence these perceptions (Shunk & Mullen, 2012), the Martin et al. study indicates that early intervention directed at supporting their academic efficacy and competence – even at the high school level – may be an important strategy for supporting engagement and reducing drop-out (Rumberger & Rotermund, 2012).

A second important practice implication of the Martin et al. (2017) study involves the positive and significant relationship that emerged between students’ motivation and their perceived motivational milieu. This motivational milieu concept is important for student engagement practice because it traces the origins of students’ task expectancies and values to forces, factors, and actors (e.g., peers and parents) that are external to the classroom and the school. For this reason, community schools (and related extended school models) merit consideration as promising practice for student engagement. As noted elsewhere, these models provide schools with an array of services and programs that, when developed and aligned appropriately, can provide a powerful engagement-focused reach into student’s peer, family, and community ecologies (e.g., Lawson. & Van Veen, 2016).

Achievement goal theory

A third line of Type 2, process-oriented engagement research can be traced to theory on student’s achievement goals. In this theory, children’s academic achievement varies according to two sources: (1) students’ achievement goals (motivation); and (2) their achievement behaviors (engagement). Following this view, students achievement-related goals are typically classified by researchers into one of three component ‘types:’ (1) task-related goals reflect student intentions to master academic tasks; (2) self-goals direct student action toward self-betterment and improvement; and, (3) other-related goals reflect student intentions to outperform others in the classroom (see also, Eccles & Wigfield, 2002).

In addition to classifying students’ achievement goals by these three types, recent research has further organized these constructs according to valence. In this view, achievement goals that are positively directed are typically called approach-oriented goals. Achievement goals that are negatively directed (i.e., those that are directed to avoid failure) are often referred to as avoidance-oriented goals. In turn, this dual stratification of student’s achievement goals by type and valence structures the 3 x 2 framework that is evident in much of the current research conversation (e.g., Conley, 2012).

Framed in this way, Cai and Liem’s (2017) study of 491 elementary school students in Singapore follows one such 3 x 2 analysis of students’ achievement goals. The purpose of this study was to understand the relationship between students’ achievement goals, motivational reasons (i.e., autonomous and controlled), and engagement. These constructs and relationships were analysed using confirmatory factor analysis and structural equation modeling.

In this study, Cai and Liem (2017) conceptualized engagement as a multi-dimensional construct consisting of affective, cognitive, and behavioral sub-components. They defined behavioral engagement as effort/persistence, cognitive engagement as elaboration; and affective engagement was defined, rather unusually, as anxiety. Each of these engagement features were measured by these researchers at multiple time-points which enabled them to account for the influence of students’ prior academic competencies on their future engagement experiences.

Results from this study indicate that when other achievement related goals are controlled statistically, only students’ self-based goals predict their cognitive and behavioral engagement. This finding is important for engagement practice because it highlights the significant cognitive and behavioral benefits that might accompany student interest in doing their own best work. At the same time, these models also revealed a potential negative consequence of students’ self-goals – namely, elevated levels of student anxiety. This latter finding, in particular, highlights needs for school psychologists to help students and parents develop goal setting practices and priorities that support students’ positive student identity perceptions and promote their short and long-term mental health (Salmela-Aro et al., 2016).

A second important practice implication that follows this study concerns the valence of school students’ achievement goals. As noted above, Cai and Liem’s study was originally designed to stratify students’ task-, self-, and other-related goals by students’ approach- and avoidance-orientations. However, the researchers did not attend to these different goal orientations in their final statistical models because of (multicollinearity) problems in their data.

In the narrative, Cai and Liem (2017) suggested that these data problems may have been caused by the difficulty young students may face in making conceptual distinctions between approach- and avoidance-oriented achievement goals. While this attribution may certainly be correct, it also ignores an important, alternative scenario. In this scenario, young children may be quite able to distinguish between approach- and avoidance-oriented goals. However, these distinctions may be obscured statistically if students’ academic motivation is jointly structured by approach- and avoidance-oriented goals.

This latter possibility is important for student engagement research and practice because it highlights an emergent research development. This development involves recent research that suggests that students’ achievement motivations may not be structured by singular goals as once theorized, but instead may reflect multiple configurations of achievement-related goals (Conley, 2012). For school psychologists and related school-based practitioners, this finding means that the pathways to students’ motivation and engagement may be far more complex and variable than what was originally proposed. This complexity highlights needs for data-driven research models that can help practitioners better attend to the variability of students’ diverse school and classroom experiences.

Type 3 engagement research: Engagement as a trait-based and identity-laden concept

Type 3 engagement research conceptualizes student engagement as an umbrella construct that subsumes motivation and other student-level variables that describe the learning process (Reschly & Christenson, 2012). In these studies, research interest typically resides in understanding how engagement may vary among different sub-populations of students (e.g., Lawson & Masyn, 2015b). This focus and interest in student engagement sub-populations has spawned a couple of important research developments. One recent example is a new line of research on ‘student engagement profiles’ (e.g., Lawson & Masyn, 2015a, 2015b; Linver, Roth, & Brooks-Gunn, 2009). Another related example is research that associates student engagement with different kinds of school and academic identities (e.g., Eccles & Wang, 2012; Goldin, Epstein, Shorr, & Warner, 2012).

Framed in this way, the Watt et al. (2017) study of 551 students in Australia contributes to a growing line of research that examines engagement as it relates to different student sub-populations and profiles (e.g., Lawson & Masyn, 2015b). In this study, the authors used indicators of student’s affective, cognitive, and behavioral engagement to estimate different profiles of mathematics engagement. Because these engagement profiles were estimated using continuous factor indicators, their analytic approach can be best described as Latent Profile Analysis (LPA) (Masyn, 2013).

Ultimately, these LPA models yielded three characteristically distinct profiles of students’ mathematics engagement. These profiles are: (1) an ‘engaged’ profile characterized by elevated levels of affective, cognitive, and behavioral engagement; (2) a ‘compliant’ profile characterized by higher levels of behavioral engagement, but lower levels of emotional and cognitive engagement; and, (3) a ‘disengaged’ profile that was characterized by low levels of cognitive, affective, and behavioral engagement.

In addition to modeling these profiles, the Watt et al. (2017) study examined each profile’s relationship to several known correlates of engagement, such as students’ perceptions of their classroom environment, as well as important demographic variables such as student age and gender. These analyses revealed that students in the ‘Engaged’ profile were typically younger, and they also enjoyed greater access to mastery-oriented classrooms and enthusiastic teachers. Meanwhile, students in the ‘Disengaged’ profile were more likely to be boys, and these students reported the lowest level of ‘school caring’ of all engagement profiles groups. Lastly, Watt and colleagues’ analysis of the demographic features of the ‘Compliant’ profile group indicated that these students were more likely to be girls than boys.

All in all, the findings from the Watt et al. (2017) study suggest that student engagement is a highly complex and contingent phenomenon. It is complex because it varies among different sub-populations of students. It is contingent because the sub-population engagement profiles examined in this study vary according to multiple developmental and contextual influences such as gender, age, and classroom characteristics. For school psychologists and other practitioners, these findings highlight needs for engagement interventions that are fit for purpose, with particular sub-populations, in particular contexts, at particular moments in time (Lawson & Van Veen, 2016). The contours and rationale for such a nuanced and integrative framework for student engagement practice and research are described in the following sections.

Implications

My charge was to highlight the practice and policy implications associated with the five empirical articles featured in the April 2017 special issue on student engagement. Together these articles signify the scholarly progress that has been made toward advancing a robust, theoretical understanding of engagement and its relationship to student motivation, learning, and instructional context. This kind of understanding-oriented research is important because it yields a set of theoretically sound, research-supported models. These models provide practitioners with specialized frameworks and language systems for describing, explaining, and predicting how engagement works across different student and school populations (Kuhn, 2012).

Granting these benefits and others, an important reminder is in order about the limitations of this overall research agenda. It provides models for practice, not models of practice. This distinction is consequential for several reasons, but one merits particular attention here. Specifically, school psychologists, educators, and other professionals who strive to foster engagement in low-income communities, often do so in environments where variations in school, family, and community resources complicate their ability to perform their daily work (Lawson & Alameda-Lawson, 2012). Faced with this persistent challenge and difficulty, these professionals may not necessarily benefit from (or want) another ‘theory’ (Filter et al., 2013; Weiss, 1995). What these professionals may want and need, however, are nuanced models of engagement practice that can help them troubleshoot their most pressing engagement-related challenges, opportunities, and concerns (Filter, Ebsen, & Dibos, 2013).

But research will not facilitate this kind of practical problem-solving and know-how unless scholars do a better job of attending to variability in their research designs. To this point, the SPI articles indicate that student engagement varies significantly by student and by situation (Martin et al., 2017; Shernoff et al. 2017), by students’ motivational style and/or orientation (Cai & Liem, 2017; Martin et al., 2017; Zhou & Rhen, 2017), by student sub-population (Watt et al., 2017), by classroom context (Shernoff et al., 2017), by motivational context and milieu (Martin et al., 2017), and even by nation state (Cai & Liem, 2017). However, despite its prominent role in explaining, qualifying, and at times confounding each study’s results, the practical relevance of this variation was largely underplayed in the five SPI engagement articles. In fact, in several cases, it was not even mentioned at all.

My critique does not minimize these articles’ important contributions to extant research and theory. My point is to highlight the difficulty school psychologists may face in deriving sound practice conclusions from research that describes engagement’s ‘general’ or ‘average’ workings, when the students they work with often present engagement-relevant strengths, needs, and challenges that are characteristically distinct from those of the ‘average’ student (Eagle et al., 2015).

Fortunately, research alternatives exist that can complement the theory building designs featured in many of the SPI articles. For example, the person-centered, statistical methods featured in the Watt et al. (2017) study offer an analytic framework that can help researchers identify sub-population differences/variations in engagement, together with their developmental and contextual influences and consequences.

Even so, the utilization of these methods alone may not create more user-friendly, practice relevant research. Methods that attend to the practical challenges posed by today’s students, schools and classroom environments are also needed.

The potential of improvement science for engagement research

For the past several years, improvement science has emerged as promising research and practice strategy for enhancing educational practices and outcomes (Bryk et al., 2015). It is based on the proposition that wholesale improvements in educational practices and outcomes, including student engagement, require the blending of two important kinds of knowledge. The first kind of knowledge is ‘basic knowledge’ – which can also be understood here as ‘theoretical knowledge’. It includes the kinds of understanding-oriented research featured in the SPI engagement articles.

The second kind of knowledge highlighted in improvement science is referred to – rather abstractly – as a ‘profound system of knowledge’. It refers to the professional knowledge and ‘know-how’ that is needed to enact theoretical knowledge within real schools and classrooms (Lewis, 2015). Examples of this ‘profound system of knowledge’ concept includes practitioner knowledge of engagement’s variation in particular schools and classrooms; their knowledge of how local systems and local politics work; their knowledge of how particular school cultures influence teaching and learning; their knowledge of student and teacher motivation, and so forth (Bryk et al., 2015).

Improvement science helps to marry these two kinds of knowledge – practical/profound and theoretical – through a deliberative set of practices, methods, and tools (Lewis, 2015). Each is designed to help practitioners and scholars execute five primary improvement-oriented goals. These goals are: (1) to identify context-specific barriers to teaching, learning, and engagement, including root causes; (2) to understand variation in classroom practices and outcomes as an important practical challenge that needs to be solved; (3) to see the system that produces the current set of student (and instructional) outcomes; (4) to use data to evaluate when change results in an actual improvement; (5) to use disciplined inquiry (by way of a process known as ‘Plan-Do-Study-Act’ cycles) to guide daily practice (Bryk et al., 2015).

These primary purposes and goals of improvement science are particularly well aligned with school psychologists’ daily work with diverse students, settings, and organizational systems (e.g. Filter et al., 2013). Moreover, when conceived in an RTI framework, improvement science affords researchers and practitioners with opportunities to gain fresh theoretical and practical insight into how learning supports might be better structured to meet student’s diverse engagement-related strengths and needs.

Several practice-focused research questions help frame this new improvement-oriented direction for student engagement research and practice. They are:

How might school psychologists and educators work together to identify students’ diverse, engagement-related strengths, needs, and challenges?

How might school professionals and researchers collaborate to develop learning environments that support the engagement-related needs of diverse student sub-populations (i.e., Tier 2 interventions)?

How can RTI teams use data to improve the design and delivery of engagement-focused interventions?

How can evidence-based interventions and other engagement-focused best practices be adapted to fit the resource needs and demands of particular students and schools?

What kinds of system-level strategies are needed to enhance the design and delivery of learning supports that foster engagement for all students?

Research structured in this way will yield new models of engagement practice, including potential contributions to theory-driven models for practice.

Conclusion

The five empirical articles featured in the April 2017 special issue of SPI contribute to understanding-oriented engagement theory and research. Known in some circles as Mode 1 knowledge generation (Nowotny et al. 2003), this now-dominant approach to engagement research provides school psychologists, educators, and other school professionals with theoretically-sound, research-supported models. These models offer important lenses for examining and describing engagement, its causal mechanisms, as well as its practical applications.

Notwithstanding the scientific value of this Mode 1 engagement research, the complexity posed by today’s diverse students and today’s global educational policy environment presents opportunities for innovative kinds of context specific, practice-embedded research (i.e., Mode 2 knowledge). Although one method for generating this kind of ‘Mode 2’ knowledge was sketched briefly in this commentary (improvement science), this kind of engagement research represents the next frontier.

Finally, it is important to emphasize that the two kinds of knowledge associated with this agenda – i.e., theoretical knowledge and practical/profound knowledge – are neither competitive nor oppositional. Both are essential, and together they provide a rich opportunity for research-and-practice partnerships. In these new partnership arrangements, the conventional ‘from research to practice’ arrangement (Mode 1) is both complemented and enriched by a ‘from practice to research’ dynamic (Mode 2). My appreciative critique has been structured in pursuit of such a mutually beneficial, bridge-building agenda for student engagement research, policy, and practice.