Children’s engagement during collaborative learning and direct instruction through the lens of participant structure

https://doi.org/10.1016/j.cedpsych.2022.102061Get rights and content

Highlights

  • Children more engaged in collaborative group (CG) than direct instruction (DI)

  • In CG, children most engaged when interacting with each other.

  • In DI, children most engaged when interacting with their teacher.

  • In both CG and DI, children least engaged when working alone.

  • CG and DI differ in bandwidth for participation, turn-taking routines, child roles.

Abstract

The joint effects of instructional approach and participant structure on children’s engagement were investigated in 24 fifth-grade classrooms during a six-week unit on wolf management. 96 four-minute episodes sampled from the lesson videos were coded in 30-second intervals for individual students’ momentary cognitive-behavioral engagement, momentary emotional engagement, and current participant structure. Multilevel ordinal regression analyses showed that children were three times more likely to be cognitively and behaviorally engaged and one-and-a-half times more likely to be emotionally engaged during collaborative groups than direct instruction. Children’s cognitive-behavioral engagement depended upon participant structure: Children in collaborative groups were most engaged in lessons during peer interaction, whereas children receiving direct instruction were more engaged when interacting with the teacher. Children were more likely to be emotionally engaged when talking with peers, the teacher, or both peers and the teacher than when alone. The study contributes to new understanding of engagement as a function of the nature and conditions for participation in lessons.

Introduction

Engagement has been identified as a key resource for students’ meaningful and successful experience in school (e.g., Christenson et al., 2012, Fredricks et al., 2004, Sinatra et al., 2015). Research on engagement during the past three decades has consistently shown its importance in students’ short-term learning and test performance, long-term patterns of school involvement and academic resilience, and development of agency and identity (Bae and Lai, 2020; Chi et al., 2018, Chi and Wylie, 2014; Nasir and Hand, 2008, Patall et al., 2019, Wang and Eccles, 2013). Lack of engagement can lead to devastating outcomes such as early school dropout. Children from underserved communities are vulnerable to becoming disengaged from schools. The underachievement of children of color including African American and Latinx children has been attributed in part to low engagement in school (e.g., Goldenberg, 2008, Johnson et al., 2001).

The good news is that engagement is malleable. There is increasing evidence that engagement is responsive to changes in the learning context, such as teachers’ instructional approach and classroom social structure (Bae and Lai, 2020, Downer et al., 2007, Herrenkohl and Guerra, 1998, Patall et al., 2019, Patchen and Smithenry, 2015). In a meta-analysis of 22 classroom studies involving students ranging from 8 to 14 years of age, Guthrie and Humenick (2004) identified four key classroom practices to increase student motivation and engagement: Emphasizing knowledge goals, allowing student choices, providing interesting reading material, and accentuating student collaboration. Unfortunately, none of these practices is common in classrooms with large enrollments of children from disadvantaged backgrounds (Arreaga-Mayer and Perdomo-Rivera, 1996, Rentner et al., 2006).

In these classrooms, the principal methods are whole class instruction featuring recitation and individual seatwork consisting of exercises in work books and skill sheets, practices that are repetitive, boring, and seldom enable high levels of intellectual involvement (Nystrand et al., 2003). McCaslin and associates (2006) completed an observational study in 145 third- to fifth-grade classrooms in 20 low-income schools. They reported that direct, teacher-led instruction predominated in virtually every classroom. Nearly 75% of instructional time in these classrooms focused on fundamental facts and basic skills, along with modest levels of elaboration and related thinking. Only 3% of instructional opportunities were devoted to higher-order thinking and reasoning. The few questions asked by students were mainly concerned with task procedures and correctness of answers; only 3% of student questions were judged to involve thinking or knowledge exploration (McCaslin et al., 2006). An even larger observational study by the National Institute of Child Health and Development (2005) in 780 third-grade classrooms confirmed that the most frequent forms of instruction were teacher-led whole group instruction and individual seatwork. By a ratio of 11 to 1 instructional activities were basic skill focused rather than involving analysis, inference, or synthesis. A caveat is these studies were done some time ago and instructional practices might have changed since then, although we are unaware of any widespread change. It is apparent that there is a need for instruction that is intellectually stimulating and personally engaging, and that the need is urgent in schools serving large numbers of poor, minority children.

The alternative to whole-class direct instruction evaluated in the current study is small-group collaborative learning. Systematic reviews of research beginning with Johnson et al., 1981, Slavin, 1983 show that collaborative learning has positive effects on achievement, variously defined depending on the studies being reviewed, in academic subjects ranging from elementary school arithmetic to college chemistry. Kyndt et al. (2013) reviewed 65 studies completed since 1995 that compared ‘face-to-face cooperative learning’ with ‘traditional instruction’ in which the teacher taught the whole class as a single group. They obtained a large effect size for achievement (ES = 0.54) and a smaller although still significant effect size for attitudes (ES = 0.15).

Despite the proven potential of collaborative learning, there is limited research that systematically compares the effects of student-centered and teacher-centered instruction on engagement during the learning process. Most of the research on collaborative learning has focused on learning outcomes (e.g., Johnson and Johnson, 2009, Nokes-Malach et al., 2015, Slavin, 2013), leaving it unclear whether and how student engagement differs between instructional approaches. Among studies that have compared aspects of student engagement between collaborative learning and direct instruction, most were short-term, cross-sectional, and usually involved small samples of students (Hänze & Berger, 2007; Nystrand and Gamoran, 1991, Patchen and Smithenry, 2015, Wu and Huang, 2007). Typically, engagement was assessed in a self-report attitude survey covering an indefinite period of time, while engagement is dynamic, and can vary from moment to moment, and lessons consist of various activities, each with a different potential to enhance or diminish engagement. To begin to fill these gaps, the present study examined student momentary engagement during lessons, taught via teacher-led whole class instruction or student-led small collaborative groups, in a multi-classroom quasi-experiment conducted in schools with large enrollments of children from low-income minority families.

Further, we widened the lens for the study of engagement during classroom lessons to include a consideration of participant structure. Participant structure entails roles for participants and expectations about the form that participation will take (Phillips, 1972). Embedded within instructional approach, participant structures can be synchronized with the “constraints and affordances” of the learning context that shape opportunities to participate (Greeno, 2015, p. 255). Given the importance of opportunities to participate for student engagement (Bae and Lai, 2020, Nasir and Hand, 2008), examining student engagement through the lens of participant structure can bring in new insights into the nature and conditions of participation in classroom lessons. Finally, to respond to the calls for diverse methodologies for studying engagement, we employed a distinctive and rigorous observational approach to examine student momentary engagement as situated in the learning context.

In the following sections, we (a) describe features of engagement during the learning process, (b) discuss the concept of participant structure, (c) compare differences in participant structures between direct instruction and collaborative learning, (d) analyze how the nature and conditions of participation influence children’s engagement, and (e) provide an overview of the current study.

Engagement refers to students’ attitude towards school and their participation in school and classroom activities. It is a meta-construct that incorporates the three components of behavioral, cognitive, and emotional involvement, according to a widely cited paper by Fredricks, Blumenfeld, and Paris (2004). Behavioral engagement refers to students’ actions and practices during academic tasks and extracurricular activities. Cognitive engagement refers to students’ investment in learning, including self-regulated strategies for learning and willingness to exert the effort to comprehend difficult ideas (Zimmerman, 1990). Emotional engagement represents a student’s affective reactions to teachers, classmates, and school activities, and sense of connectedness to and valuing of school (Skinner et al., 2008). Other dimensions of engagement that have been proposed include agentic engagement, which represents students taking initiative (Jang et al., 2016), and collaborative engagement (Järvelä et al., 2016) or social-behavioral engagement (Linnenbrink-Garcia et al., 2011), which point to the shared nature of engagement during collaborative learning. The multidimensional conception of engagement enables a rich characterization of children’s involvement in school; however, in practice it can be difficult to distinguish among the dimensions, such as between cognitive and behavioral engagement (Eccles, 2016).

Extending Fredricks et al.’s conceptualization, Pekrun and Linnenbrink-Garcia (2012) postulated cognitive-behavioral engagement as a dimension separate from purely cognitive or purely behavioral engagement. Cognitive-behavioral engagement refers to cognitive processes that are initiated by the learner intentionally, which may include problem-solving, self-regulated learning, and use of cognitive and metacognitive strategies in learning. Informed by this line of thinking, in this study we examined cognitive-behavioral engagement as an integrated form of engagement during classroom learning activities. We also examined children’s emotional engagement as reflected in their affective experience during learning activities. Enjoyment and excitement during lessons are associated with increased attention, use of deep learning strategies, and higher achievement whereas boredom and frustration are associated with decreased attention, use of shallow learning strategies, and decreased achievement (Pekrun et al., 2017).

The view of engagement as concrete, effortful, and detectable from fine-grained analysis of behavior while students learn is the foundation of the ICAP framework (Chi & Wylie, 2014; Chi et al., 2018), where students’ engagement is identified from overt behavior and differentiated into one of four modes: Interactive (dialoguing), Constructive (generating), Active (manipulating), and Passive (receiving). Drawn from the literature that associates student learning outcomes with different levels and kinds of engagement, the ICAP framework embodies the theory that higher levels of cognitive engagement (I > C > A > P) are associated with better learning outcomes, conceptual change in science, and better performance on standardized measures of achievement (Chi and Wylie, 2014, Miller et al., 2014, Wang and Holcombe, 2010). However, in the data set that we analyzed there was often insufficient information to distinguish between the Interactive, Constructive, and Active modes, and it seemed prudent not to force the distinctions.

In this study, engagement was viewed in terms of students’ participation in classroom lessons, organized in terms of participant structures (Philips, 1972; also called participation structures, Au, 1980; or participation frames, Goffman, 1981). A participant structure encompasses who the participants are, the relative authority of the participants, their roles and their rights and obligations, and expectations about the form that participation will take (Cornelius & Herrenkohl, 2004). Expectations may be codified as explicit rules (raise your hand to get a speaking turn) or reflected in implicit norms (overlapping speech is okay if you are building on the speaker’s contribution; avert your eyes to show respect when speaking to an authority figure or, conversely, look people straight in the eye to show sincerity). Participant structure is not fixed, but may evolve over time as dialogue progresses, varies as a function of the age, status, and cultural background of the participants, depends on changing relationships among the participants, and varies with the stage or phase of the activity in which the participants are engaged.

Participant structures have cultural roots. Pioneering research involving Native American children (Phillips, 1972) and Native Hawaiian children (Au, 1980, Au and Mason, 1981) concluded that children thrive when structures for classroom talk resemble the forms of participation common in the home and community. When the teacher allowed children to take open turns and perform jointly with peers, as well as participate in small-group activities with no status hierarchy, the interactional rights and responsibilities of the students and the teacher were more closely balanced, leading to greater engagement. Similar results have been observed in research with African American and Latinx students, where collaborative group work has been associated with improved achievement, enhanced attitudes, and increased pro-social behavior (Van Ryzin et al., 2020, Watkins, 2002, Webb and Farivar, 1994).

Power and authority are parameters of participant structures. Tabak and Baumgartner (2004) addressed the symmetry of power and balance between authoritative and persuasive discourse. They examined three participant structures in which the teacher’s role varied. Compared to teacher as monitor and teacher as mentor, the structure of teacher as partner was more effective in fostering students’ appropriation of modes of scientific inquiry, as the teacher shared the authority to talk with students and achieved a balance between the “authoritative discourse” and “internally persuasive discourse” (Bakhtin, 1981).

Although few studies have examined students’ engagement in lessons through the lens of participant structures (Chinn et al., 2001; Herrenkohl & Guerra, 1998; Patchen and Smithenry, 2015, Wu and Huang, 2007), recent research suggests that how participation is structured impacts engagement, and the impact may differ among individual students across phases of lessons that involve different instructional practices (Webb et al., 2018). To further understand this impact, we next compare participant structures within direct instruction and collaborative learning.

Direct instruction and collaborative learning assume different modes for participating in lessons and varied distribution of control over lesson activities (Chinn et al., 2001, Green et al., 1988, Wu and Huang, 2007). Each instructional approach has an expected or canonical participant structure, and norms that encourage, promote, or privilege certain types of participation and social interaction, and thus provide students with distinctive opportunities for engagement.

Direct instruction is a teacher-centered instructional approach in which teachers exercise primary control over classroom talk. According to Baumann (1984), “In direct instruction, the teacher … tells, shows, models, demonstrates, teaches the skill to be learned. The key word here is teacher, for it is the teacher who is in command of the learning situation and leads the lesson … (p. 287).”.

Direct instruction is valued for its efficiency in knowledge transmission, which according to an influential line of research arises because skillful direct instruction enables teachers to keep students’ cognitive load within manageable limits (e.g., Kirschner et al., 2006). In learning complex materials, teachers can ensure students have fewer encounters with misconceptions and provide corrective feedback as soon as they detect a misconception. Because of its presumed efficiency and effectiveness, direct instruction is favored over collaborative group work in schools with large enrollments of nonmainstream children (Arreaga-Mayer and Perdomo-Rivera, 1996, Rentner et al., 2006).

In direct instruction, the teacher governs most aspects of lessons and students have limited control over when they can speak, little or no say about the topic, and negligible authority to evaluate whether contributions to a discussion are acceptable (Wells & Arauz, 2006). Students’ opportunities to ask questions, redirect the topic, or otherwise take initiative are circumscribed; when they talk, their speech often consists of only a word or a phrase (Cazden, 2001, Nystrand et al., 2003).

The canonical participant structure in direct instruction is students interacting with the teacher. However, there will also be times when students interact with each other, whether authorized by the teacher or not, and times when students work alone, reading silently or completing individual exercises. As direct instruction privileges teacher-student interaction, it is likely that students’ engagement can reach the highest level when they interact directly with the teacher.

In collaborative learning, students work together in small groups to help one another learn (cf. O'Donnell, 2006). Collaborative groups feature distribution of power and responsibility among the teacher and the students (Anderson et al., 1998, Cornelius and Herrenkohl, 2004). As the instructional paradigm centers on students, the teacher is expected to delegate authority and become a facilitator of learning, while students are expected to manage both the problem space and the social relationship space while working together (Barron, 2003, Sun et al., 2015). The exercise of authority is more fluid since both students and the teacher are expected to contribute to the process (Zhang et al., 2018). Students speak freely without raising their hands or waiting for the teacher to call on them, negotiating turn taking among themselves. In successful collaborative groups, students both initiate their own thinking and serve as a responsive audience for the contributions of other group members (Herrenkohl & Guerra, 1998).

In some structured approaches to collaborative learning, students are pre-assigned roles such as scriber (Johnson et al., 2013). In less structured approaches, such as Collaborative Reasoning (Anderson et al., 1998), the approach used in the present project, roles are not prescribed nor is front loaded social skills training provided. Nonetheless, one or two child leaders naturally emerge in most groups, who facilitate turn taking, help the group stay on track, and monitor progress toward goals (Li et al., 2007). Emergent leaders tend to talk more than other members of the group, ideally without dominating the discussion, although they are sometimes regarded as bossy by other children. Sun et al. (2017) found that children were able to employ the leadership skills they had acquired during Collaborative Reasoning discussions in a cooperative problem-solving activity with a new group of peers with whom they had not worked before. Children who had experience with Collaborative Reasoning achieved better solutions to a spatial reasoning problem and had more positive attitudes toward the problem-solving experience.

Collaborative learning is valued as a means to attain higher-order cognitive goals. Research has shown that when managed well by the teacher and child leaders, collaborative learning can lead to better comprehension of text (Murphy et al., 2018), greater conceptual growth (Howe and Zachariou, 2019, Larrain et al., 2021), and improvement in skills of argument (Reznitskaya et al., 2009). However, when not managed well, groups can suffer from lack of focus, experience negative emotions such as frustration with other group members, and fail to complete group tasks (Barron, 2003).

While there will be periods during collaborative learning when students interact with the teacher or work alone, the canonical participant structure of collaborative groups is students interacting with each other. The approach encourages children to talk to and work with peers, share ideas and resources to learn, and complete tasks together. Collaborative groups thus provide children with plenty of chances to communicate and the freedom to use ways of speaking that they are comfortable with from home and community, which may be particularly important for African American and Latinx children to be engaged in classroom lessons (Emdin, 2011).

The theory that guides the present study is that student engagement depends on the nature and conditions of participation afforded by the instructional approach and realized in participant structures. The two instructional approaches compared in this study, direct instruction and collaborative learning, differ in the bandwidth for participation, expectations about the process of participation, and the roles available to participants.

The bandwidth for participation is more limited in direct instruction than collaborative learning. Ordinarily one student at a time ‘takes the floor’ when nominated by the teacher, during teacher-led whole class instruction in episodes framed of the canonical participant structure, students interacting with the teacher. Access to the floor must be divided between the teacher and all of the students. Opportunities for speaking expand when a class is split into small groups because time is divided among fewer participants. During most teacher-led discussions, time is apportioned among the teacher and the students. Teachers have preeminent rights to speak, often take much of the time, and may express more than half the words spoken during discussions (Cazden, 2001, Cheng et al., 2015). In collaborative small-group discussions among peers, teacher time is nil and almost all of the time is available for student speaking turns.

Previous research is consistent with the premise that the bandwidth for participation is greater in collaborative learning than direct instruction. Chinn et al. (2001) compared teacher-directed discussions with student-managed collaborative discussions with the same small groups of students, controlling for the difficulty and interest of the stories the students read and discussed. They found that student words per minute were almost twice as high during collaborative discussions. Teachers substantially reduced their own talk during collaborative discussions, but less teacher talk was not the only reason for the increase in rate of student talk. Students actually talked faster in collaborative discussions. Moreover, there were large qualitative shifts in students’ talk between the two forms of discussion. As compared to student talk during teacher-directed discussions, student talk during collaborative discussions more frequently involved elaborating text propositions, making predictions, using evidence, and expressing alternative perspectives.

In many teacher-led whole class discussions, not everyone who wants to speak will be able to get the floor during surges of enthusiasm, which occur often when children are exploring a question interesting to them, such as whether a pack of wolves should be killed. Our conjecture is that the energy available during surges is diminished or lost if it is not realized in participation. Thus, during these moments, the direct instruction teacher is in the position of damping down enthusiasm and failing to capitalize on the teachable moment, or relaxing rules for participation and permitting call outs and side conversations. The common practice of choral responding may be seen as an accommodation of direct instruction to harness motivational energy.

Besides differences in bandwidth, many features of the process of participation shift between teacher-led discussions and peer-managed discussions. Typically, in teacher-directed whole-class discussions, teachers steer the discussion with strings of questions in repeated sequences of teacher initiation, student response, and teacher evaluation or feedback, the ever-present IRE/F pattern (Sinclair & Coulthard, 1975). Research establishes that the vast majority of teacher questions during whole class discussions are focused, rather than open ended, that student answers are usually brief and unelaborated, and that teacher feedback is mostly simple and unelaborated (e.g., Molinari et al., 2013, Nystrand et al., 2003). The IRE/F pattern has been primarily a script for recitation rather than a frame that enables productive dialogue and, accordingly, the pattern has been repeatedly criticized as ineffective for developing students’ thinking (Greeno, 2015).

A variety of modifications have been proposed to help teachers employing direct instruction improve the thoughtfulness of talk in their classrooms, including asking higher-order questions to stimulate thinking (Morris & Chi, 2020), increasing wait time (pausing after a question or pausing after the student response) so that students have more opportunity to think (Rowe, 1986), and revoicing students’ answers to extend their thinking (O’Connor & Michaels, 2019). However, it has proved challenging to make widespread or large-scale changes in the nature of classroom talk, as “the kinds of shifts in practice called for … require that one change deeply ingrained patterns of talk, are complex, and require an apprenticeship (for both teachers and students) into new ways of participating in discussions” (O’Connor & Michaels, 2019, p. 173).

Finally, the roles that students play during lessons vary distinctively between direct instruction and collaborative learning, with implications for students’ understanding of themselves as passive learners and bystanders or knowledge seekers and thinkers. A teacher-led whole class discussion consists of series of short conversations between the teacher and one or a few students. During a conversational episode, the remaining students are positioned as bystanders, who may overhear the conversation but do not otherwise participate (Clarke, 2015, Goffman, 1981). Students frequently in the bystander role are undoubtedly less likely to sustain engagement than students who get roles as speaker or addressee (Chen et al., 2021). Teacher concern about the disengagement of bystanders may be part of the reason classroom talk is resistant to change. If they see heads turning away and bodies beginning to squirm, teachers can be expected to reduce the complexity of questions, maintain a brisk pace instead of prolonging wait times, and cut back on elaborated feedback or follow up questions, in order to make speaking turns available to bystanders starting to disengage and get through the lesson without the class falling into disorder. Students in small collaborative discussions can be bystanders, too; the difference is that they are less likely to be forced into the role because of the communication frame.

During teacher-directed whole class discussions, students cannot be sure when they will get a turn for speaking or what question they will be expected to answer. Our conjecture is that unpredictability about speaking turns is an impediment to the extended thinking that undergirds active or intense engagement. Students have to wait to hear the question before they start formulating an answer. In an open collaborative discussion, students are in a better position to engage in focused thinking and initiate speech production before a turn for speaking since they have personal control over when they speak and what they speak about. This may enable a faster rate of speaking, more complete and developed thinking, as well as a higher level of engagement. Not knowing when they will have turns may especially undermine participation of English language learners who cannot as readily as native speakers call up words and arrange them in an appropriate syntactic frame (Skehan, 2009).

The teacher is the primary audience for students during direct instruction and the teacher can be a problematic audience. Students are well aware that teachers believe they know the answers to most of the questions they ask, so students recognize that ordinarily instructional questions are not genuine requests for information. They are probably aware that instructional requests for arguments ordinarily are not genuine, either, as teachers usually believe they already know the reasoning and the conclusion that should follow and are not actually open to being persuaded. Teachers do not act as true parties to arguments. They say, “Good reasoning,” not, “You convinced me to change my mind.” Thus, a discussion in a direct instruction classroom may be perceived as an exercise in performing well rather than an opportunity for mastering knowledge, solving a problem, or reaching the best decision. Motivational research establishes that a performance orientation, as opposed to a mastery orientation, can have a pernicious effect on engagement (Ames, 1992, Dweck, 1986, Rogat and Linnenbrink-Garcia, 2019, Rolland, 2012).

Other students are the audience in peer-managed discussions. Unlike addressing the all-knowing teacher, speakers will recognize that like themselves other students may be uncertain or lack information; thus, explanations or informational statements offered to them usually are genuine. That doesn’t mean every statement is meant to be taken literally, of course, because students use humor, irony, and devices such as rhetorical questions. Importantly, the arguments students have with each other are usually genuine, in the sense that the arguments are advanced to inform and persuade.

Our theory is that the roles students play in lessons influence their understanding of themselves as knowledge seekers and thinkers. In direct instruction, teachers do most of the initiating, explaining and evaluating, leaving students with circumscribed opportunities for extended reasoning and, we suppose, little cause to believe that their own thinking counts for much. Thus, direct instruction students are positioned as followers of the teacher’s reasoning. In collaborative groups, as they were realized in this study, students had the chance to become active seekers after knowledge, arguers and decision makers who engaged in chains of reasoning themselves. Motivational research indicates that teaching that supports student autonomy and sense of agency or self-determination enhances engagement and has many other benefits (Ryan and Deci, 2000, Stroet et al., 2013) whereas teaching that is highly controlling is demotivating and associated with fear of failure, contingent self-worth, and avoidance of challenge (Bartholomew et al., 2018).

Morris et al. (2018) sought to verify the foregoing account of who does the reasoning during teacher-led discussions as compared to student-led discussions by analyzing a low-inference indicator of reasoning, the distribution of words used to indicate how ideas are connected. Analysis of video clips systematically sampled from lesson videos, the same set of videos that the current study is based upon, showed that students participating in student-managed small-group discussions used the connectives because, so, if, then, and, and but at more than four times the rate of students participating in teacher-managed whole-class discussions. In contrast, interestingly, direct instruction teachers used the set of connectives at over twice the rate of collaborative learning teachers. The picture that emerges of direct instruction is one in which the teacher does most of the reasoning, with students filling in small pieces when requested. Whereas, in collaborative learning students do most of the reasoning, with the teacher occasionally redirecting when students go off track. As a specific indicator of whether students considered themselves to be active decision-making agents, Morris et al. (2018) searched the video clips for the performative verb phrases I think and I know. Students in collaborative groups used these phrases at twice the rate of direct instruction students.

We hypothesize that students’ engagement will be highest when they participate through the canonical participant structure for that instructional approach. The reasoning is that expectations for participation will then be synchronized with the affordances and constraints of instruction. For example, consider the case of student–student interaction during a direct instruction lesson. It is more likely than student–teacher talk to be off topic; even when on topic, there may be tension over student–student talk, and it may be disallowed by the teacher for violating rules of participation.

The present study was designed to evaluate the theory that instructional approaches and participant structures jointly influence children’s engagement. Students in direct instruction and collaborative groups classes studied the same custom-made socio-scientific curriculum unit for six weeks. Students can develop skills in reasoning and perspective taking, as well as obtain discipline specific knowledge, when they engage in socio-scientific inquiry (Sadler et al., 2007).

Our major specific hypothesis was that children in collaborative groups would be more engaged than children receiving direct instruction. As detailed in the previous section, the rationale for this prediction is that, as compared to collaborative groups, direct instruction affords children fewer opportunities to participate, that typical features of participation during direct instruction interfere with engagement, and that in direct instruction children are often constrained to roles that do not afford the highest levels of engagement. Our second specific hypothesis was that, when participating in lessons conducted according to a certain instructional approach, students’ engagement will be highest when they participate through the canonical participant structure, because the mode of participation will be synchronized with the affordances and constraints of instruction.

Students’ cognitive-behavioral engagement and emotional engagement were assessed by coding episodes sampled from videos of lessons throughout the socio-scientific unit. Using observational measures allowed us to capture momentary engagement in the changing contexts in which individual students were situated (Sinatra et al., 2015) and overcome other shortcomings of retrospective self-report measures (Greene, 2015). Multilevel ordinal regression analyses were employed to account for the ordinal and hierarchical nature (momentary observations nested within individual students nested within classrooms) of engagement coding.

To understand the impact of participant structure on individual student’s engagement, we tracked each child’s momentary interaction patterns during the lessons analyzed in this study. Different from the typical approach, where participant structure is assumed to be fixed over time and constant throughout a classroom, we took a dynamic approach at the individual student level to capture changing interaction patterns that may vary within and across students and change over the course of a lesson. We examined four types of overarching interaction patterns indexed by who a student is interacting with: student interacts with peers, student interacts with the teacher, student interacts with both peers and the teacher, and student is alone not interacting with others.

The prediction that students in collaborative groups will be more engaged than students experiencing direct instruction is of legitimate empirical interest. Many educators and policy makers will judge that, far from heightened engagement, students in lightly supervised small groups will squabble among themselves, get drawn into fruitless disputes about procedure, waste time doing nothing while waiting for the teacher to rescue them from one difficulty or another, sit looking out the window while a couple of students do the work, and go off task to discuss extraneous topics, instead of focusing on the academic tasks set for them. The current study aimed to provide empirical evidence addressing these common beliefs and to offer practical implications for engaging instruction.

Section snippets

Participants

Included in the present study were 149 fifth-grade students (64 boys, 85 girls) from 24 classrooms in low-income Midwest schools who participated in this project across two academic years. Each year, 18 classrooms were recruited, nine of which had a predominant enrollment of African American students, and the other nine a predominant enrollment of Latinx children. Depending on the school, from 78.6% to 98.8% of the students were registered for free or reduced-priced lunch. Classrooms within

Results

Results are organized in three sections that present descriptive statistics about engagement, multi-level ordinal logistic regression analyses of factors that influence engagement, and video recorded excerpts from CG and DI lessons to serve as concrete examples of the findings from the statistical analysis.

Discussion

This study is one of the first to examine through fine-grained observational measures how student momentary engagement differs between collaborative learning and direct instruction. The findings from the study decisively confirm our primary hypothesis that students experiencing collaborative groups (CG) would be more engaged than their peers receiving direct instruction (DI). Comparing any two children, one a CG student, the other a DI student, the odds that the CG student would be behaviorally

Limitations and future directions

Several limitations need to be acknowledged. Participant structures were aggregated into superordinate categories in the present study. A more detailed and refined analysis of participant structures (cf. Au, 1980, Tabak and Baumgartner, 2004) may well yield further insights into the conditions that stimulate engagement. In the present study, a moment was operationally defined as 30 s. Here, too, a more fine-grained analysis may yield further insights; for example, Imai et al. (1992) showed that

Implications and conclusion

Despite its limitations, this study has several noteworthy findings with theoretical and practical implications. The major finding was that higher levels of engagement are much more likely during peer-managed collaborative learning than during teacher-managed direct instruction. The theory proposed to explain this finding is that bandwidth for participation is more limited during direct instruction than collaborative learning, that turn-taking routines during direct instruction interfere with

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The research reported in this article was supported by the Institute of Education Sciences, U.S. Department of Education, through Grant R305A080347 to the University of Illinois at Urbana-Champaign, Richard C. Anderson, Principal Investigator. The opinions expressed are those of the authors and do not represent views of the

References (104)

  • B.W. Miller et al.

    The effects of reading to prepare for argumentative discussion on cognitive engagement and conceptual growth

    Learning and Instruction

    (2014)
  • J.A. Morris et al.

    Instructional discourse and argumentative writing

    International Journal of Educational Research

    (2018)
  • C. O’Connor et al.

    Supporting teachers in taking up productive talk moves: The long road to professional learning at scale

    International Journal of Educational Research

    (2019)
  • E.A. Patall et al.

    Agency and high school science students’ motivation, engagement, and classroom support experiences

    Journal of Applied Developmental Psychology

    (2019)
  • K. Stroet et al.

    Effects of need supportive teaching on early adolescents’ motivation and engagement: A review of the literature

    Educational Research Review

    (2013)
  • M.-T. Wang et al.

    School context, achievement motivation, and academic engagement: A longitudinal study of school engagement using a multidimensional perspective

    Learning and Instruction

    (2013)
  • R.C. Anderson et al.

    Intellectually stimulating story discussions

  • J. Andrade

    Repetition and dual coding in procedural multimedia presentations

    Applied Cognitive Psychology

    (2010)
  • C. Ames

    Achievement goals and the classroom motivational climate

  • C. Arreaga-Mayer et al.

    Ecobehavioral analysis of instruction for at risk language-minority students

    The Elementary School Journal

    (1996)
  • K.-H.-P. Au

    Participation structures in a reading lesson with Hawaiian children: Analysis of a culturally appropriate instructional event

    Anthropology and Education Quarterly

    (1980)
  • K.-H.-P. Au et al.

    Social organizational factors in learning of read: The balance of rights hypothesis

    Reading Research Quarterly

    (1981)
  • C.L. Bae et al.

    Opportunities to participate in science learning and student engagement: A mixed methods approach to examining person and context factors

    Journal of Educational Psychology

    (2020)
  • M.M. Bakhtin

    The dialogic imagination: Four essays by M.M

    (1981)
  • J.F. Baumann

    The effectiveness of a direct instruction paradigm for teaching main idea comprehension

    Reading Research Quarterly

    (1984)
  • B. Barron

    When smart groups fail

    Journal of the Learning Sciences

    (2003)
  • C.T. Butts

    Social network analysis with sna

    Journal of Statistical Software

    (2008)
  • C.B. Cazden

    Classroom discourse: The language of teaching and learning

    (2001)
  • J. Chen et al.

    The role of friendships in shy students’ dialogue patterns during small group discussions

    Contemporary Educational Psychology

    (2021)
  • J. Chen et al.

    Triple alignment: Congruency of perceived preschool classroom social networks among teachers, children, and researchers

    Frontiers in Psychology

    (2020)
  • Y. Cheng et al.

    Moving from recitation to open-format literature discussion in Chinese classrooms

    Instructional Science

    (2015)
  • M.T.H. Chi et al.

    Translating the ICAP theory of cognitive engagement into practice

    Cognitive Science

    (2018)
  • M.T.H. Chi et al.

    The ICAP framework: Linking cognitive engagement to active learning outcomes

    Educational Psychologist

    (2014)
  • C.A. Chinn et al.

    Patterns of discourse in two kinds of literature discussion

    Reading Research Quarterly

    (2001)
  • S. Clarke

    The right to speak

  • L.L. Cornelius et al.

    Power in the classroom: How the classroom environment shapes students’ relationships with each other and with concepts

    Cognition and Instruction

    (2004)
  • J.T. Downer et al.

    How do classroom conditions and children's risk for school problems contribute to children's behavioral engagement in learning?

    School Psychology Review

    (2007)
  • C.S. Dweck

    Motivational processes affecting learning

    American Psychologist

    (1986)
  • C. Emdin

    Dimensions of communication in urban science education: Interactions and transactions

    Science Education

    (2011)
  • J.A. Fredricks et al.

    School engagement: Potential of the concept, state of the evidence

    Review of Educational Research

    (2004)
  • S.D. Gest et al.

    A Reputation for success (or failure): The association of peer academic reputations with academic self-concept, effort, and performance across the upper elementary grades

    Developmental Psychology

    (2008)
  • E. Goffman

    Forms of talk

    (1981)
  • C. Goldenberg

    Improving achievement for English language learners

  • J.L. Green et al.

    Lesson construction and student participation: A sociolinguistic analysis

  • B.A. Greene

    Measuring cognitive engagement with self-report scales: Reflections from over 20 years of research

    Educational Psychologist

    (2015)
  • J.G. Greeno

    Classroom talk sequences and learning

  • J.T. Guthrie et al.

    Motivating students to read: Evidence for classroom practices that increase reading motivation and achievement

  • L.R. Herrenkohl et al.

    Participant structures, scientific discourse, and student engagement in fourth grade

    Cognition and Instruction

    (1998)
  • M. Imai et al.

    Properties of attention during reading lessons

    Journal of Educational Psychology

    (1992)
  • Cited by (0)

    View full text