Adolescents' engagement profiles and their association with academic performance and situational engagement

Highlights

• Four profile groups were identified based on students' overall engagement.

• Profile groups differed with students' academic GPAs.

• Differences in students' overall engagement were linked to situational engagement in actual learning situations.

Abstract

This study examined adolescents' engagement profiles and their association with situational engagement and academic performance in lower secondary school settings. The data consisted of 301 Grade 7 students from Finland. By utilizing person-oriented approach with Latent Profiling Analysis on subscales capturing students' overall engagement, four subgroups of students with different overall engagement profiles were identified. These profile groups were further analyzed with respect to differences in student background (gender and maternal education), academic performance (GPA, and reading and mathematics tests), and lesson-specific situational engagement. The profile groups showed differences with respect to gender, maternal education, and GPAs as well as situational engagement. The findings provided new insights into the associations between overall and situational engagement by documenting concordance but also disconcordance between them, and by showing situational fluctuation of engagement in particular among students with high overall engagement.

Introduction

The predictive association between student engagement and academic learning has been widely acknowledged in the literature (Fredricks et al., 2004; Wang and Eccles, 2011). Nevertheless, prior studies have limitations because they have mostly approached engagement by focusing on students' typical engagement experiences over time (Eccles & Wang, 2012), here termed as overall engagement. While the situational fluctuation of student engagement has been evidenced in prior literature (e.g., Martin et al., 2015), attempts to capture situational variation in students' engagement via repeated measures in authentic learning situations (i.e., situational engagement assessed by self-ratings of students in situ) are still relatively rare. Although prior literature provides some theoretical propositions on how overall and situational engagement are connected (e.g., Lawson & Lawson, 2013), empirical analyses on this relationship are needed. Studies utilizing a person-oriented approach have identified subgroups of students with respect to engagement (Lawson & Masyn, 2015b; Wang & Peck, 2013), but information on situational engagement of students with varying profiles of overall engagement is lacking. This kind of information would be useful for identifying and supporting students with low engagement, and for providing teachers' insight into their students' situational engagement in the classroom and from lesson to lesson. In order to bridge the gap between the two separate strands of engagement literature, the present study set out to examine and increase understanding on the association between overall and situational engagement. A person-oriented approach is utilized to identify subgroups of students with similar patterns of overall engagement, and these subgroups are subsequently analyzed with respect to their situational engagement, academic performance and background factors.

Student engagement is a multidimensional construct that combines different aspects related to students' commitment and involvement with school and learning (e.g., Appleton et al., 2008). According to the widely acknowledged conceptualization by Fredricks et al. (2004) student engagement consists of three distinct, yet interrelated, components: behavioral engagement, emotional engagement, and cognitive engagement (see also Fredricks, Ye, et al., 2019; Wang et al., 2011). In the literature, the term behavioral engagement is typically used to refer to students' active involvement and effort focused on learning and academic tasks, positive conduct (e.g., completing assignments and preparing for classes), and participation in extracurricular activities (Fredricks et al., 2004; Wang et al., 2011). Emotional engagement is seen to encompass students' sentiments toward school and learning, and a sense of belonging with respect to school and its people (e.g., Appleton et al., 2006; Fredricks et al., 2004). Cognitive engagement captures students' cognitive investment in learning and schooling. Drawing from two somewhat distinct lines of literature, it focuses both on self-regulated strategies, which students use to master knowledge or skills, and on the value they place for schoolwork and its importance for their future (e.g., Appleton et al., 2008; Fredricks et al., 2004).

While the tripartite conceptualization presented above continues to be the most prevalent definition for the construct of student engagement (Fredricks, Reschly, & Christenson, 2019), also additional components, such as social engagement (Fredricks, Wang, et al., 2016) or agentic engagement (Reeve, 2013) have more recently been introduced into the conceptualization of student engagement. These somewhat varying definitions of what constitutes the core of student engagement can be considered to reflect the evolving nature of the field. In the similar vein, boundaries between dimensions of engagement are not necessarily clear cut for interlinking constructs as, for example, student effort can be seen to function as an indicator of behavioral as well as cognitive engagement (e.g., Fredricks, Filsecker, & Lawson, 2016). In addition, an agreement is yet to be reached on whether indicators and facilitators of engagement (the latter construed, e.g., as competence experiences) should be distinguished and treated as separate to engagement (see e.g., Fredricks, Reschly, & Christenson, 2019). Another example of the continual evolution in the field is the conceptualization of the construct of disengagement, which first was construed as the absence of engagement, but has lately been viewed as an independent concept including maladaptive behaviors and emotions that cannot be explained always as low engagement (Fredricks, Ye, et al., 2019; Lawson, 2017; Skinner et al., 2009).

Despite of the differences in definitions and point of views among scholars, there is a wide agreement that engagement is a multidimensional construct, which critically adds to understanding of the whole of student behaviors, feelings, and cognitions within learning-related contexts. Engagement is also recognized as being malleable, which means that it is responsive to external support, and, therefore, it may be accommodated via interventions (Lawson & Lawson, 2013; Wang, Decol, & Henry, 2019).

The present study is founded on the stance that student engagement needs to be approached on two levels: via students' self-reports of their behaviors, emotions, and cognitions captured in situ in actual learning-related situations (i.e., situational engagement) as well as via students' self-appraisals of their typical engagement with school and learning over time (i.e., overall engagement). The vast majority of prior studies have approached engagement by using measures that aggregate information across time from different situations and time points, and by aiming to capture students' average experiences (Eccles & Wang, 2012). These studies have provided valuable knowledge on the association between engagement and academic performance (e.g., Ladd & Dinella, 2009; Wang & Holcombe, 2010), and documented change in engagement longitudinally over time (Archambault et al., 2009). They have also provided information on the role that student characteristics play in engagement by documenting lower engagement among boys than girls (e.g., Lam et al., 2016), and by indicating that family's low socio-economic status may pose a risk for lower engagement (Li & Lerner, 2011; Linnankylä & Malin, 2008). In addition, they have increased our understanding on how engagement is not uniformly formed, but instead sub-populations with several different engagement dispositions can actually be recognized among students (Lawson, 2017; Lawson and Masyn, 2015a, Lawson and Masyn, 2015b). However, when engagement is approached at its overall level, situation-specific variations in students' learning-related experiences are not acknowledged and captured (cf. Eccles & Wang, 2012). Thus, knowledge on student engagement that can be gained by focusing on such variations is needed to move the field forward.

While only limited number of studies have focused on situational engagement with repeated assessments capturing students' experiences in authentic learning situations, those have consistently documented intra-individual variations in students' situational engagement (Martin et al., 2015; Pöysä et al., 2018; Vasalampi et al., 2016). These studies have also begun to accumulate evidence on factors that may have an influence on fluctuation in situational engagement. Recent studies have shown, for example, that situational engagement is higher in lessons of non-academic subjects than in academic subjects (Pöysä et al., 2018; Shernoff et al., 2003), and that situational engagement is associated with emotional support in teacher–student interactions during the lesson (Pöysä et al., 2018). Thus, engagement is not only malleable over time, but fluctuates from one learning situation to another as well. Nevertheless, studies capturing both overall and situational engagement simultaneously are scarce, and investigations that would provide empirical evidence on the relations between the two, are still needed.

Despite the yet limited body of empirical findings on the relations between overall and situational engagement, assumptions about the conceptual link have been put forward in previous literature. In their prominent review of student engagement research and theory Lawson and Lawson (2013) laid out a persuasive framework according to which students' various acts of engagement that occur at particular moments in time (i.e., at the situational level), stem from students' prior dispositions of engagement as well as chancing external conditions (see also Lawson, 2017). This notion of engagement dispositions does not, however, imply that these dispositions should be seen as inherently stable or deterministic (Lawson & Lawson, 2013). Rather, engagement dispositions are malleable and, they can, therefore, be molded, for example, by experiences gained at particular moments in time (Lawson & Lawson, 2013). Aligning with this proposition, the present study advocates the view that overall engagement (cf. engagement dispositions) and situational engagement (cf. acts of engagement) are transactionally related. Investigating the two constructs simultaneously can throw light, first, on whether the relation would be similar for students with different levels of overall engagement, and second, what kind of factors are associated with shaping of students' engagement.

In the present study, the tripartite definition of student engagement formed the basis for assessment at both overall and situational level, but some differences were inevitable with respect to operationalizations. First, when assessing students' overall engagement, the dimensions of behavioral, emotional, and cognitive engagement were retained as independent from one another (Appleton et al., 2006; Wellborn & Connell, 1987) whereas with respect to situational engagement, behavioral and cognitive engagement showed substantial association in the students' self-ratings (e.g., attending to a task, sustained cognitive problem solving and aim for mastery are likely to be highly linked), which is why in effect they were merged (see Vasalampi et al., 2016). Second, situationally bound facilitators (i.e., students' competence experiences and help-seeking) were included in self-ratings at situational level while those were not separately assessed at overall level. The theoretical rationale for including competence experiences as situational facilitators draws from the self-system motivational model (Connell & Wellborn, 1991). Help-seeking, in turn, was conceptualized as a facilitator of re-engagement in interaction and learning tasks during the lesson (Marchand & Skinner, 2007). Finally, at the situational level, disengagement was seen as a relevant aspect of learning situations and in line with prior literature (cf. Skinner et al., 2009) the term disaffection was used and defined as maladaptive behaviors and emotions occurring within the lesson.

Previous studies examining student engagement have largely been conducted using traditional variable-oriented designs, which focus on universal and linear associations between variables in a particular sample (see Bergman & Andersson, 2010). While such studies provide important insights into student engagement, they do not consider the potential of subgroups of individuals with different profiles or patterns of engagement. Thus, variable-oriented designs are likely to tap into the average student but may remain silent on actual individuals (Bergman & Trost, 2006; Fredricks, Ye, et al., 2019; Lawson & Masyn, 2015a). By focusing on patterns of similarities and differences within the measured indicators, person-oriented approach can provide novel insights on subgroups of individuals who engage in different ways (Bergman & Trost, 2006; Eccles, 2016; Lawson & Lawson, 2013). For example, by revealing different engagement dispositions with person-oriented approach, Lawson and Masyn (2015b) were able further analyse the differences these subgroups had in their educational attainment outcomes. In the present study, person-oriented approach was used for its potential for bringing light to fluctuations of situational engagement which may be different for different subgroups of students and have different associations with individual factors (gender and maternal education), and academic performance factors.

As conceptualizations and operationalizations of engagement are not uniform in research field, the prior engagement literature that utilizes a person-oriented approach show a broad diversity. Some studies, such as that by Wang and Peck (2013), have identified subgroups with different profiles of engagement using all three dimensions of engagement (i.e., behavioral, emotional, and cognitive) simultaneously, but some studies have focused only on some specific dimensions (e.g., Li & Lerner, 2011). The number of identified subgroups also differs, and sometimes only three patterns of engagement have been identified (e.g., Watt et al., 2017) whereas in some studies, the number of identified subgroups have been higher, for example six, or seven (e.g., Lawson & Masyn, 2015b; Archambault et al., 2009, respectively). Nevertheless, studies utilizing person-oriented approach to examine engagement through repeated ratings of situational engagement are exceptionally rare (for an exception, see e.g., Schmidt et al., 2017). However, despite of diversity in the used variables and the number of identified subgroups, the prior findings can be seen to lead into a conclusion that more advanced knowledge of engagement can only be achieved by paying attention to within-group variations and to profiles of individuals who are engaged in different ways (e.g., Lawson & Masyn, 2015b; Schmidt et al., 2017). Such knowledge would critically enhance understanding on the role that engagement has in learning processes, and help to devise tools, tailored interventions, and guidelines for practitioners to support students in day-to-day practice (cf. Lawson, 2017).

The present study was set to reach a more comprehensive understanding of students' overall and situational engagement in lower secondary school settings. While evidence has accumulated on student engagement in general, studies assessing engagement with repeated measures at the situational level in contextually bound learning situations are much more rare (Eccles & Wang, 2012). To our knowledge hardly any studies have set out examine overall as well as situational engagement simultaneously in order gain insight on variation of situational engagement in different subgroups of students. Although the present data and measures partly overlap with prior studies of the authors on intra-individual situational variation in students' engagement, and correlations between overall and situational engagement (Pöysä et al., 2018, Pöysä et al., 2019; Vasalampi et al., 2016), the present study adds a unique contribution by utilizing the person-oriented approach for examining students' overall and situational engagement. The following research questions were set:

• What kind of subgroups of overall engagement can be identified among Grade 7 students? Based on previous findings showing that engagement may vary among different subgroups of students (e.g., Lawson & Masyn, 2015b; Watt et al., 2017), it was expected (Hypothesis 1) that several distinct subgroups would be identified.

• To what extent do the identified subgroups differ in a) students' background factors (gender and maternal education), and b) academic performance (GPAs, and tests in reading and mathematics)? First, in line with earlier literature indicating higher engagement for girls than for boys (e.g., Lam et al., 2016), and lower engagement for students from family's with lower socioeconomic statuses (e.g., Linnankylä & Malin, 2008), it was expected that subgroups would differ based on the distribution of students' background factors (Hypothesis 2a). Second, based on previous findings (e.g., Ladd & Dinella, 2009), it was expected that subgroups with higher overall engagement would manifest higher academic performance, whereas subgroups with lower engagement would manifest lower academic performance (Hypothesis 2b).

• To what extent do the identified subgroups differ in a) mean levels of situational engagement (behavioral/cognitive engagement, emotional engagement, disaffection, competence experiences, and help-seeking), and b) variation of situational engagement across lessons? As prior literature provides theoretical propositions on links between overall and situational engagement (Lawson & Lawson, 2013), it was expected (Hypothesis 3) that subgroups would differ on mean levels of students' situational engagement. Due to the lack of prior studies, no hypothesis was set for variation in situational engagement by subgroup (RQ 3b).