Abstract
Adolescents’ peer networks tend to segregate by relative language proficiency, but students from all linguistic backgrounds benefit academically from classroom peer relationships both within and across English learner (EL) and non-EL classified groups. We drew upon social network analysis of student survey data in 46 English and math middle school classrooms and qualitative analysis of a subset of these classrooms (N = 10) to address the following: (a) How do demographics differ in classrooms with more or less academic peer network linguistic integration? and (b) How do teachers’ classroom practices relate to differences in the linguistic integration of students’ academic peer networks? Findings from this analysis add to the literature on the complex relationships between classroom characteristics, linguistic integration, and teacher practices.
As classrooms in the United States become increasingly diverse along linguistic, ethnic/racial, cultural, socioeconomic, and other lines—and continuing educational inequities along those dimensions become increasingly well documented (O’Day & Smith, 2016)—the pedagogies and principles underlying educator practices that can support equitable and high-quality educational experiences for all children have never been more important. In this study, we turn to one aspect of such variation that has a strong impact on the education of a sizable portion of K-12 students in the United States: the diversity among students who are at various points in their development of English as an additional language. For the purposes of this analysis, we examine classrooms that contain both students who are classified in their schools as learners of English (English learners or emergent bilinguals, whom we call ELs) and those who are either English-only users or bi/multilingual students classified in schools as fluent speakers of English (whom we call non-ELs). We acknowledge the complexity of socio-institutional mechanisms that label and categorize students according to levels of language development as well as the linguistic diversity within each of the categories we have chosen to employ (e.g., Kibler & Valdés, 2016). However, given broad educational concerns about inequitable educational opportunities provided to students classified as English learners, we contend that the focus of this analysis remains a critical one. Research suggests that students’ personal social networks tend to segregate by relative language proficiency, as we discuss further below, but students from all linguistic backgrounds benefit academically from classroom peer relationships both within and across EL and non-EL language groups (Molloy Elreda, Kibler, Johnson, & Karam, 2018).
We employed a mixed-methods approach, drawing upon social network analysis of student survey data as well as qualitative analysis of classroom observational field notes, in order to address the following research questions: (a) How do classroom demographics differ in those classrooms with more or less academic peer network linguistic integration? and (b) How do teachers’ classroom practices relate to differences in the linguistic integration of students’ academic peer networks? To answer these questions, we first determined classrooms’ levels of academic peer network integration (the opposite of segregation) by language status (EL vs. non-EL)—what we define as peer network linguistic integration—in 46 English language arts and math middle school classrooms (Grades 6–8), using the Freeman Segregation Index (Freeman, 1972). We then identified the demographic characteristics and shared and unique teaching practices found in a subset of classrooms (N = 10) identified as having either low/decreasing peer network linguistic integration or high/increasing peer network linguistic integration. Findings from this analysis add to the literature on the ways in which teaching practices may relate to positive youth relationships and dynamics in today’s linguistically diverse classrooms.
Theoretical Framing
To understand the classroom as a dynamic developmental context for EL-classified youth and their peers, this paper uses a developmental systems framework informed by García-Coll et al.’s (1996) work on minority youth development alongside ecological theories of language development (Kramsch, 2002; van Lier, 2002, 2004). In considering particular issues facing these youth as linguistically minoritized members of classrooms, 1 García-Coll et al.’s (1996) integrative model of minority youth development specifies that schools are among the many social settings in which youth are exposed to and engage with discrimination, both structural and interpersonal (see also Marks, Ejesi, McCullough, & García-Coll, 2015). We recognize that these dynamics affect EL-classified youth and conceptualize more linguistically integrated classrooms as potentially more likely to feature supportive rather than discriminatory interpersonal experiences for these students. In relation to language specifically, we draw upon ecological theories of language development that conceptualize language-learning environments such as classrooms in terms of the dynamic interplay between language users and the resources available to them, prominent among them teachers and peers (Kramsch, 2002). Language development is thus an interactive process, not simply a matter of internal cognition (Leather & van Dam, 2003). As a result, it is vital to better understand the interactive opportunities afforded—or inhibited—through both teacher-student and peer relationships (van Lier, 2002, 2004).
Developmental systems frameworks emphasize dynamic and mutually influential layers within an individual (e.g., cognitive, emotional), a given social context (e.g., interpersonal interactions, social settings), and sociopolitical structures (e.g., public education, immigration law, structural racism). In examining the linguistic integration of classroom ecologies, we attend explicitly to dynamic systems at two of those three levels: the individual and the social context. In doing so, we focus on both the teacher-student relationships that affect students’ individual sense of belonging in a classroom as well as the peer relationships that shape the creation of a sense of community among students in a classroom social context, while also acknowledging that these processes are inseparable from the larger sociopolitical structures that have shaped them. A dual emphasis on both the individual and the social context has guided our investigation of the literature as well as our data: as a result, the sections that follow explicitly attend to both individual belonging and classroom community.
We also conceptualize classrooms as dynamic rather than static and seek to understand the mutually influential relationships among classroom demographics, teacher practices, and peer linguistic integration within a classroom. Teacher practices—which we define as observable behaviors and actions teachers undertake during lessons—both create and are created by the classroom ecology (J. Cohen & Grossman, 2016). 2 In this sense, teacher practices do not exist in a vacuum: they are responsive to the actions of students in front of them. At the same time, however, teacher practices have the potential to influence not only a students’ sense of belonging in a classroom (Chhuon & Wallace, 2014; Meeuwisse, Severiens, & Born, 2010) but also the nature of students’ relationships with peers (Farmer, McAuliffe Lines, & Hamm, 2011). Because relations among these processes may vary for different individuals and in different groups of individuals (e.g., classrooms; Lerner, Phelps, Forman, & Bowers, 2009), we focus on several classrooms of interest and assess similarities and differences in patterns observed.
Literature Review
Peer Language Segregation Among Adolescents
Several studies have documented adolescents’ tendencies toward forming peer groups, or social networks, that segregate according to their dominant or home languages, especially when youths’ use of an additional language is infrequent, their proficiency in that language is in its early stages, or immigration is recent (e.g., Aboud & Sankar, 2007; Hamm, Brown, & Heck, 2005; Titzmann & Silbereisen, 2009). Institutional mechanisms in U.S. schools that sort students into courses or academic tracks according to assessed language proficiency often further limit students’ access to peers who are linguistically different from themselves (e.g., Kibler & Valdés, 2016; Valdés, 2001). And while social and academic connections with other language learners can provide space for learning and shared struggle without “competition from native speakers” (García & Bartlett, 2007, p. 6), interaction with English-dominant or bi/multilingual peers is also important, both as a key resource for access and exposure to how English is used in the context of U.S. education (e.g., Carhill-Poza, 2011, 2015; Hawkins, 2004; Valdés, 2004), and as a source of connectedness to classmates that facilitates greater school engagement (e.g., Benner, 2011; Tsai, 2006). Because cross-language peer relationships often also cross racial/ethnic boundaries, findings from this body of research are relevant to the current study, and there is evidence that relationships both within and across racial/ethnic groups provide unique benefits to all involved. For instance, while same race/ethnicity friendships bring important opportunities for identity development and psychological well-being (Kiang & Fuligni, 2009; Tatum, 2017), cross race/ethnicity friendships have also been found to promote better psychological well-being (Bagci, Rutland, Kumashiro, Smith, & Blumberg, 2014; Faircloth & Hamm, 2011) and better academic outcomes (Kawabata & Crick, 2015; Mickelson, 2015) across racial/ethnic groups. Similarly, previous analyses of the current dataset have demonstrated that same- and cross-language-status academic relationships (EL/EL and EL/non-EL) significantly contributed to across-year growth in academic engagement for all students (Molloy Elreda et al., 2018).
Classrooms as Sites for Linguistically Integrated Adolescent Peer Relationships
Classrooms in which EL-classified students are enrolled alongside English-only or bi/multilingual peers have the potential to provide important opportunities for all students to develop a range of both same- and cross-group relationships with peers, but such classrooms are linguistically and socially complex spaces. Foundational to any growth in peer relationships is students’ sense that they belong as legitimate members of their classroom community: they must believe that “they matter, and that they are recognized, understood, and included” in their instructional setting (Markus, 2008, p. 89). Yet students of color are less likely than White students to experience such validation or sense of belonging in school (Landsman & Lewis, 2006; Markus, 2008; Olsen, 1997). In the case of most students classified as ELs, they belong to not only minoritized racial and ethnic groups but linguistically minoritized groups as well. In fact, monolingual (English-only) expectations are so pervasive in the education system that they are frequently unrecognized or left unchallenged (Valdés, Capitelli, & Alvarez, 2011), and as a result, multilingual and multicultural competencies are framed in a deficit perspective (K. D. Gutiérrez & Orellana, 2006; Reyes & Ervin-Tripp, 2010; Rymes, 2010). Unsurprisingly, research has documented a trend in which many students classified as ELs have been “intimidated by the sociolinguistic environment” (Harklau, 1994, p. 263) of secondary classrooms in which non-EL-classified students are also enrolled and participate less often in such classrooms than in EL-only settings (Duff, 2001, 2002, 2004; Harklau, 1994; Yoon, 2010). However, research has also affirmed the role of teachers in creating classroom environments that more actively support secondary EL-classified students’ sense of belonging. For example, in the same studies where students reported that they felt socioemotionally safer in English as a second language (ESL) classrooms, Harklau (1994) and Yoon (2010) found that ESL teachers would explicitly discuss American cultural norms, explore students’ cultural backgrounds, and build relationships with individual youth. Although Harklau (1994) and Yoon (2010) found such practices to be more likely to happen in ESL classrooms, other researchers have documented effective practices by secondary teachers in classrooms serving both EL- and non-EL-classified students as well. Pass and Mantero (2009) found, for example, that even a practice as commonplace as greeting students by name helped students feel more supported in linguistically diverse classrooms with non-ELs. In addition, students classified as ELs have been found to develop positive relationships with their teachers and report feeling academically supported when teachers actively engaged with students’ out-of-class lives and grew to know their families (Michael, Andrade, & Bartlett, 2007).
Beyond creating an environment in which EL-classified students have a sense of legitimacy and belonging in the classroom, teachers can also influence the development of the larger classroom community through facilitating their engagement with peers. Peer interactions are critical sites of learning for ELs, as research has shown that working closely with peers, either in pairs or in small groups, can benefit students academically and linguistically (Baker, Gersten, & Lee, 2002; Calderón, Slavin, & Sánchez, 2011; Cole, 2013; Gersten et al., 2007; Pyle, Pyle, Lignugaris/Kraft, Duran, & Akers, 2017). When Brooks and Thurston (2010) observed middle school students in different configurations of instruction—including whole group, small group, and one-on-one instruction—they reported that EL-classified students were more likely to participate verbally while interacting with a group of peers than with the whole class. On the other hand, Verplaetse (1998) has documented that when middle school content teachers led whole-class discussions rather than assigning group work, teachers asked fewer open-ended and high-level cognitive questions to students classified as ELs than their non-EL peers and sometimes even completed their answers, which limited opportunities to engage in extended discourse. In tandem with language production, EL-classified students’ content understanding has been shown to improve in middle school science and high school social studies classes when teachers established cooperative learning groups (Buck, Mast, Ehlers, & Franklin, 2005; Szpara & Ahmad, 2007).
Specifically, research has demonstrated that when students whose English development is emergent work with peers at more advanced levels of proficiency, they have opportunities for increased participation and content understanding in secondary classrooms (Bunch, 2006; R. Gutiérrez, 2002; Lotan, 2008; Pon, Goldstein, & Schecter, 2003; Short, 1994; Wassell, Martin, & Scantlebury, 2013; Yoon, 2007). However, without the teacher’s intentional grouping of linguistically heterogeneous (mixed EL and non-EL) students, these students may confront what E. G. Cohen and Lotan (1995, 1997, 2014) call “status problems.” According to this idea, students have different statuses depending on peers’ perception of their academic standing, social standing, or societal standing at large, and these factors influence group work dynamics. Such effects can be seen at the outset of group work with the very selection of partners: Yoon (2010) found that when students were free to choose their partners, EL-classified students were not approached for collaboration by non-ELs and had difficulty finding partners. Similarly, Sharkey and Layzer (2000) found that students interacted more in high school classrooms where teachers used seating charts to encourage interaction across different linguistic backgrounds. When students are allowed to choose their own seating, what often results can be a clustering that “replicates the general positioning of marginalized groups in the US” (McKay & Wong, 1996, p. 584). Once students are grouped, low-status students do not talk as much as others, and when they do, their contributions may be ignored by other members of the group (E. G. Cohen & Lotan, 2014), trends that have important implications for students classified as ELs.
However, research has also shown that teachers can mitigate status problems and facilitate productive group work with linguistically diverse students specifically by establishing norms around group work (Bunch, Lotan, Valdés, & Cohen, 2005; Khisty, 1993), creating tasks that require complex problem solving skills and multiple abilities (Bunch, 2014; Bunch et al., 2005), requiring individual accountability as well as interdependence (Jacob, Rottenberg, Patrick, & Wheeler, 1996), asking open-ended questions for students to discuss with peers (Daniel, Martin-Beltran, Peercy, & Silverman, 2016), providing evaluation criteria for judging group products (E. G. Cohen, Lotan, Abram, Scarloss, & Schultz, 2002), praising low-status students with public and valid feedback (E. G. Cohen & Lotan, 1995), and assigning or delegating responsibilities to students (Bunch, 2006; Lotan, 2008). When teachers implement these practices, EL-classified students have greater opportunities to engage in contextually rich language use and to give as well as receive academic assistance (Jacob et al., 1996). Teacher practices that facilitate effective collaboration among linguistically heterogeneous students thus can allow otherwise marginalized students to learn more, can contribute to counteracting prevalent prejudices, and can ultimately support improved individual and group performance (E. G. Cohen & Lotan, 2014).
Just as some teacher practices have been observed to promote a sense of community for students classified as ELs, other practices have actively undermined it. Research has suggested that the way in which a teacher addresses topics of cultural and linguistic diversity in heterogeneous classrooms can strongly influence the positioning of EL-classified students’ identities and ultimately, their sense of community among peers. In particular, studies have shown that talking about issues of diversity is not in itself sufficient. For example, when teachers initiated academic discussions about race or multiculturalism without recognizing and confronting the racism that students faced in their own lives, classroom hierarchies were maintained (Talmy, 2010). Harklau (2003) likewise found that when high school teachers attempted to create “color-blind” spaces, students classified as ELs developed identities that were “seemingly unmarked” (p. 87), but they were not provided with the means to “explore the xenophobia and discrimination they had encountered” (p. 95).
Method
We addressed our first research question by using social network analysis to explore whether the end-of-year level and across-year growth in peer network linguistic integration in a classroom were associated with classroom demographic characteristics. To answer the second question, we then used social network analysis to select classrooms at different levels of linguistic integration for qualitative analysis to compare patterns identified in teachers’ practices for creating individual belonging and building classroom community.
Research Context
The present study is part of a larger mixed methods project that included 46 classrooms in two middle schools in one school district in a state located in the South-Atlantic region of the United States. Of the 46 total classrooms, maximum variation sampling (Patton, 2015) was used to select 13 classrooms for qualitative study in efforts to represent a range of content areas, grade levels, varied levels of EL population density, and instructional quality (Kibler et al., in press). For the current study, social network analysis was used to quantify the extent to which each classroom’s academic peer relationships were segregated or integrated by EL status, a process which allowed us to identify 10 of these 13 classrooms as having either low/decreasing or high/increasing linguistic integration. Qualitative analysis of observational field notes then allowed us to examine patterns of specific teacher practices that were present in those more or less linguistically integrated classrooms.
The research team for the current project consisted of lead researchers who were senior and emerging scholars with expertise in the teaching and learning of EL-classified students, peer and youth-adult relationships, social network analysis, and quantitative/qualitative/mixed methods inquiry. Data collection and analysis were undertaken by these scholars as well as graduate research assistants, all of whom were former teachers seeking advanced degrees in education and extensively trained in quantitative and qualitative research methods, particularly in taking and coding field notes. Through weekly research team meetings, lead researchers provided initial and ongoing training on the use of observational methods, participated as classroom researchers, and conducted data analysis jointly with graduate research assistants.
Participants
Adams and Braxton Middle Schools, located in Hampshire County School District, 3 both had established policies through which EL-classified students of varied proficiency levels were placed in linguistically heterogeneous classrooms. Demographics from the first year of data collection, which are very similar to those in subsequent years of the study, are presented in Table 1, showing that Adams’s and Braxton’s populations generally mirrored state and local averages in terms of EL-classified populations, economic disadvantage, and racial/ethnic composition, although with somewhat fewer Hispanic/Latino students than are enrolled nationally. (We use Hispanic/Latino and other racial/ethnic labels because these were the demographic categories used in the publicly available sources from which we drew.)
National, State, School, and Classroom Demographics
Note. AA = African American, W = White, F = female, M = male; ELP = English language proficiency; ELA = English language arts. All data but ELP scores rounded to nearest whole number.
Race/ethnicity data retrieved from federal, state, and local websites using their terminology.
Economic disadvantage measured by enrollment in free/reduced-price school meals program, the only data available at the time of the study.
Assessment score range was 200–600, with 400 indicating “proficient.”
There were three academic tracks with placement through teacher recommendations and assessments: Basic (B, lowest track, often enrolling special education students via an inclusion model), Standard (S, middle track), and Honors (H, most advanced track).
In the larger set of 46 classrooms, we examined data from 736 students (53% female) across the 23 math and 23 English language arts (ELA) classrooms (23 sixth-grade classrooms, 11 seventh-grade classrooms, and 12 eighth-grade classrooms) over three academic years (fall 2013 to spring 2016). Classrooms had an average size of 19 students. Twenty-nine percent (n = 212) of participating students in the study were classified as ELs (based on scores on the WIDA ACCESS™ [WIDA], an English language proficiency assessment administered annually in 31 states across the country) and had an average WIDA proficiency level of 4.18 (on a scale of 1–6, 4 range = 1.8–5.7) in the spring prior to the start of the study year. Reclassification from EL-status in this state depended on scoring at a 5.0 or higher overall and in Literacy on the WIDA assessment. EL-classified students in our sample included those who scored above a 5.0 overall (see Table 1), but they did not also score a 5.0 in Literacy. As a result, they were considered EL-classified by the state. Reclassified students were not considered EL-classified students in our analyses and were therefore not included in the average WIDA score noted here. (They comprised only 35 of the 736 students in our study.) The district kept records of reclassified students for only the 2 years mandated by federal law, and as a result, we were not able to gather further information on “ever-EL” students (Umansky, Thompson, & Díaz, 2017) for purposes of demographics or analysis.
The students in these 46 classrooms were racially/ethnically diverse, with 30.7% of students self-identifying as Hispanic/Latino, 29.1% as White, 21.3% as African American, and 18.8% as being of mixed or another race/ethnicity. Ninety-four percent of all EL-classified students self-identified as students of color. Among students in this full sample, 30 languages (including English) were spoken. Information on the demographic characteristics of each of the 10 classrooms used for our qualitative analysis is included in Table 1.
We did not have access to course schedules for students in our study, but both schools typically placed EL-classified students at WIDA proficiency levels 1 to 3 in separate ESL classes instead of English language arts classes. Students of all proficiency levels were enrolled in math, science, or social studies classes, where depending on staffing and scheduling, they may have had an ESL teacher or teaching assistant in addition to the content-area teacher. The placement of ESL specialists and their particular roles in classrooms varied considerably, but their presence or absence is included in Table 1 for the 10 classrooms selected for qualitative analysis. 5
Data Collection
Demographic Surveys and School District Data
We gathered a range of demographic data on students and classrooms via student surveys and data compiled by the school district, including students’ age, race/ethnicity, EL status, languages spoken at home, standardized test score performance, and assessed English proficiency levels. At the classroom level, data included class subject, academic track, grade level, and mobility (i.e., number of students moving in and out of the classroom during the year).
Network Surveys
We collected information about students’ peer relationships in the classroom via social network surveys (items drawn or adapted from Moody, Brynildsen, Osgood, Feinberg, & Gest, 2011; Ryan & Shim, 2012) administered in the fall and spring. In these surveys, students were asked questions to capture both friendship and academic relationships. The item used to capture academic helping relationships in the classroom was one in which students were asked to identify the classmates with whom they talk when “trying to get work done in this class.” These types of peer nomination-based items have a long history in investigations of patterns of relationships in classroom peer networks (e.g., Cairns, 1983; Moreno, 1934), and have been used extensively in studies of classroom friendships and help-seeking relationships among middle school students (e.g., Moody et al., 2011; Ryan & Shim, 2012) and in studies of cross-race/ethnicity and cross-language peer relationships (e.g., Aboud & Sankar, 2007; Carhill-Poza, 2011; Goza & Ryabov, 2009; Hamm et al., 2005). Because we hypothesized that teacher practices are more likely to have a stronger influence on students’ academic rather than their friendship networks, we focused our analysis on the former. Selection of a single measure is common in social network analysis, at least in part because there is no standard conceptual rationale or procedure for aggregating nominations across multiple items or over time (including academic help-seeking peer nominations, e.g., Ryan & Shim, 2012; Smith & Peterson, 2007). Furthermore, we follow traditions in social network analysis of utilizing incoming ties as a reliable approach (Zemljič & Hlebec, 2005), a choice that is further bolstered by the logic of the question we asked, in that academic helping relationships (i.e., the students you talk to when “trying to get work done in this class”), unlike friendships more generally, may be useful to both participants without necessarily being reciprocal (e.g., Bowman-Perrott et al., 2013; Smith & Peterson, 2007).
For purposes of this analysis, these peer nominations were used to quantify how linguistically segregated each classroom was—in other words, the extent to which academic relationships were segregated by EL status (or, conversely, how linguistically integrated they were across language learner status boundaries). Language-based network segregation was operationalized by the Freeman Segregation Index (Freeman, 1972), a commonly used index (e.g., Faris & Felmlee, 2011; Gest, Davidson, Rulison, Moody, & Welsh, 2007; Moody, 2001) for comparing the number of observed cross-group ties to the number randomly expected by chance. This index accounts for the composition of a classroom, in particular the opportunities for cross-group ties available in a setting (e.g., the number of EL or non-EL classified students present in a class). The value was calculated for each classroom as one minus the observed proportion of cross-language-status ties divided by the proportion of randomly expected cross-language-status ties, if EL status had no bearing on nominations. The resulting index falls on a scale from 0 (a classroom network that is not segregated at all), to 1 (a fully segregated classroom network with no cross-language-status ties). For instance, a value of 0.40 would reflect a classroom network in which cross-language-status ties are reduced from their random expectation by 40%.
For the present analysis, we focused on classrooms that showed the most and least fall to spring growth in segregation (spring score minus fall score) and on classrooms that were most and least segregated by the end of the school year. Specifically, we selected from the classrooms we investigated qualitatively those that were one standard deviation (SD) or more either above or below the mean of the full sample (N = 46) in spring segregation (spring segregation index M = 0.21, SD = 0.18, ranging in our full sample from 0 to 0.68) or across-year growth in segregation (M = −0.06, SD = 0.18, ranging in our full sample from −0.39 to 0.44). As mentioned above, 10 of the 13 classrooms in our study fit one of these criteria, with five that were high (or increasing) in segregation, and five that were low (or decreasing) in segregation.
It is conceptually and rhetorically important to focus on the outcome supported by research and theory—linguistic integration—rather than its opposite (segregation), although we recognize that they are two ends of the same continuum. As a result, in the findings below we use the term integration and define low/decreasing integration classrooms as those with either higher segregation scores in the spring or higher growth in segregation over the course of the school year. Similarly, we define high/increasing integration classrooms as those with either lower segregation scores in the spring or less growth in segregation during the year. 6
Observational Protocol
The research team employed an observational protocol developed during the piloting phase of the larger project to allow observers to keep running records of interactions between teachers and students as well as among peers. We observed each classroom between eight and 10 times over eight months, during which we audio- and video-recorded the classes while taking extensive field notes. Field researchers—including the first author and graduate students—were trained to record the events of the classroom as a whole while also paying special attention during observations to individual teacher-student and peer interactions. Each researcher observed one classroom throughout an entire school year to build rapport and familiarity with the teachers and students but did not otherwise have any past or present relationships with the schools or their teachers. Because of the linguistic diversity in any class, it was not possible for researchers to have expertise in all languages spoken. However, all researchers conducting observations spoke at least one non-English language and had bicultural expertise, having grown up or lived abroad for multiple years. In all possible instances, researchers were placed in classrooms in which they spoke the most prominent non-English languages used by students in the classroom. Researchers also consulted with speakers of the non-English languages they did not know when those languages were used by students in the classroom; audio and video recordings were reviewed with those consultants to then add relevant detail to fieldnotes.
Data Analysis
We used a combination of inductive and deductive qualitative data analysis approaches in our study (Fereday & Muir-Cochrane, 2006). The inductive process began with the team’s re-reading and discussion of approximately half of the field note documents for each teacher in the study. This process produced a full, detailed thematic set of teacher practices inductively grouped into five categories (Creating Individual Belonging, Creating a Classroom Community, Setting Up Productive Class Routines, Developing Academic and Linguistic Expertise, and Engaging with Student Discourse). We iteratively revised this initial draft of the codebook through subsequent coding as well as insights and findings from our review of the literature. In cases where a set of practices identified in the literature was not yet included in the coding scheme, we added codes as needed. Using the first iteration of our codebook, researchers (who had conducted classroom observations and written field notes for at least one of the 10 classrooms) used the web-based software Dedoose™ to code all field notes, applying the five main categories of teacher practices to the notes. Initial reliability was established through double-coding and team discussions using approximately 10% of the data before researchers coded the remainder of the notes individually. The codebook was again revised as needed to accommodate all practices found in the data and clarify relationships and distinctions among them.
For the purposes of our analysis relating to classroom linguistic integration, we chose to focus on two of the five teacher practice categories (Creating Individual Belonging and Creating a Classroom Community), hypothesizing that these two types of practices were most directly relevant to classroom integration. Research team members then used the same process described above to complete axial coding (Marshall & Rossman, 2011) of the specific teacher practices found in these two categories. Coders also completed reliability tests for each set of axial codes, and the codebook was updated as needed based on the results. All excerpts were double-coded, and any disagreements or questions were resolved through weekly meetings in which the full set of coders (three in total) reached agreement through a consensus process (Hill et al., 2005). The full research team (including noncoders) was consulted as needed to complete this consensus process if coders could not do so as a group. Through re-reading field note excerpts, we determined two key sets of practices across teachers that appeared consistently in the data related to individual belonging (offering praise or validation; connecting with students socially) and three sets related to classroom community (facilitating interactive peer learning; providing behavioral support; addressing diversity and discrimination). All steps of this data analysis process were completed blind to integration group: in other words, coders did not have any knowledge of which classrooms were either low/decreasing or high/increasing.
We then used the results of the social network analysis to divide teachers into low/decreasing and high/increasing integration groups and then re-read field note excerpts in order to determine which teacher practices were found across both sets of classrooms, as well as which were found uniquely in only one or the other. A practice did not need to be present in all classrooms to be included: However, singular instances that appeared only once in the data were either grouped into larger relevant practices or not included in the findings. In this way, we relied on recurring patterns but also attended to the ways in which individual practices may be linked.
In addition to the reliability tests mentioned above, we employed several other methods to ensure the validity of our analysis. One of these was the consensus process described above. Another was the inductive/deductive processes through which we aligned the teacher practices emerging in our field notes with those observed and reported on in extant literature to generate our coding structure. An additional way that we sought to ensure validity was via an informal auditing process that we followed during coding, in which two researchers on the team oversaw the coding as it progressed so as to check assumptions and pose relevant questions from a more distant perspective.
Findings
To assess how low/decreasing and high/increasing linguistic integration classrooms differed from each other demographically and how teachers’ classroom practices relate to differences in the linguistic integration of students’ academic peer networks, we first describe patterns in classroom demographic characteristics before then presenting the thematic findings regarding teacher practices.
Demographic Characteristics of Low/Decreasing and High/Increasing Integration Classrooms
In recognition of the potential for demographic characteristics of classrooms to influence teacher practices, and to assess the extent to which those demographic characteristics might help to explain differences in peer linguistic integration, we present information about classrooms in our full sample (N = 46) at different levels of classroom integration and growth/decline in integration (see Table 2).
Demographics by Peer Network Linguistic Integration, Full Sample (N = 46)
Note. EP = English proficiency; ELA = English language arts. Independent-samples t tests for high versus low and decreasing versus increasing suggest no significant differences between groups on percentage EL, average state assessment scores, average English proficiency scores and range, languages in addition to English, or mobility. Analyses of variance and t tests comparing integration and integration growth across grade levels, academic tracks, and subject areas also show no significant differences between any of these groups.
Decreasing includes “0.”
EP score range was calculated by subtracting the maximum from the minimum WIDA score in each classroom. We also tested the WIDA standard deviation, minimum, maximum, and median WIDA scores per classroom, and none were significant.
Two separate sets of t tests—one set comparing low versus high integration classrooms and another set comparing decreasing versus increasing integration classrooms—were computed for all classroom-level demographic characteristics in which continuous variables were available (percentage EL-classified population; average state assessment scores in the content area of the class, either ELA or math; average and range of English proficiency scores as measured by WIDA; number of languages spoken in the class in addition to English; and student mobility). In addition, analyses variance were computed to assess whether level or growth/decline of integration (treated as continuous variables) systematically varied by grade level (sixth, seventh, and eighth) or by academic track (remedial, standard, and honors); an additional t-test was computed to assess whether level or growth/decline of integration systematically varied by subject area (English language arts vs. Math). Across all tests, we found no statistically significant associations between demographic characteristics and level or growth/decline of integration.
Teacher Practices in Low/Decreasing and High/Increasing Integration Classrooms
Although low/decreasing and high/increasing integration classrooms were not distinguishable by demographic characteristics, they did differ according to teacher practices along the dimensions of: offering praise or validation, connecting with students socially, facilitating interactive learning among peers, providing behavioral support, and directly addressing issues of diversity and discrimination. Findings for each are presented in Table 3 and below, with practices found in both sets of classrooms followed by practices unique to the low/decreasing and high/increasing integration classrooms, respectively.
Teacher Practices in Classrooms at Different Levels of Peer Network Linguistic Integration
Note. EL = English learner; L/D = low/decreasing; H/I = high/increasing.
Total number of excerpts in a category.
Numbers in parentheses relate to frequency of a shared practice in low/decreasing versus high/increasing integration classes.
All percentages relate to total number of excerpts in a given category (e.g., 172 excerpts comprise 61.9% of all 278 “Offering praise and validation” excerpts).
Offering Praise or Validation
The first key set of teacher practices related to offering praise or positive feedback to students and validating students in relation to their work, effort, or class participation. We found fewer instances overall in low/decreasing integration classrooms (77, or 27.7% of the total) than in high/increasing integration classrooms (201, or 72.3% of the total), encompassing both shared and unique practices. 7 Furthermore, those found only in low/decreasing contexts were relatively rare (8.6%), while those unique to high/increasing settings were more frequent, accounting for more than one-quarter of the total instances overall (see Table 3.)
Praise and validation in both sets of classrooms
Teachers across both sets of classrooms validated students’ efforts to contribute verbally to whole-group discussions and, especially in ELA classrooms, in response to open-ended questions asked of students. They also praised students individually for quality, progress, or completion of work and at times used physicality (i.e., a high five or pat on the back) to reinforce these messages. Teachers praised students for their willingness or eagerness to share or participate in whole-class settings, and some praise came in the form of announcing or acknowledging student accomplishments in ways that elicited peer recognition, such as holding up a student’s behavior or work as a model for peers. Beyond these common patterns, however, differences in praise and validation practices were identified between the low/decreasing integration and high/increasing integration classrooms.
Praise and validation in low/decreasing integration classrooms
Praise and validation in low/decreasing integration classrooms occurred in both individual and group contexts but tended to be directed toward alleviating academic anxiety, emphasizing a mind-set of fixed individual strengths, or conveying differential expectations and rewards for students depending on their EL or non-EL classification. First, in efforts to mitigate students’ expressions of academic anxiety, teachers both reassured students about grades and encouraged effort either after students voiced frustration or during a difficult or time-consuming task. For example, Ms. Miller announced to her class after a test in which many students did poorly, “I know it is frustrating to get a grade you don’t want, but sometimes the best way to learn is by making mistakes.” While such comments validated students’ efforts, they were reactive, functioning to address the problem of student anxiety or frustration rather than proactively promote student confidence and self-efficacy. Next, praise or validation that commented on students’ performance also tended to emphasize students’ individual strengths as stable rather than dynamic, reflecting a fixed mind-set (Dweck, 2006): Ms. Bayley, for example, often complimented students on being “smart” but rarely commented on students’ dedication or the hard work they undertook to arrive at a correct answer.
We lastly identified a pattern through which teachers communicated differential expectations, and differential praise rewards, to students based on language status. Non-EL-classified students tended to be praised and validated through encouragement to speak or present in front of the class, to share work with peers, or to expand on their ideas. Students classified as ELs, on the other hand, were praised for being prepared with school supplies, making nonverbal contributions (even when students were at advanced levels of English proficiency), or having tangential academic skills (e.g., computer coding). They also received validation in the form of being told by teachers of their expectations for students to make high grades. Although this final example appears to hold EL-classified students to high academic standards, the other patterns position them as expected to do the minimum necessary to be successful. Conversely, students classified as non-ELs were positioned not only as effective students in their own right but also as leaders with the potential to contribute to the learning of their peers.
Praise and validation in high/increasing integration classrooms
Teachers in high/increasing integration classrooms were unique in using praise or validation to emphasize the learning process over inherent ability. Teachers additionally followed up on praising students by asking further questions or asking for clarification, showing that doing something well (and receiving praise) was not an end-point in itself, but could be integrated into the process of continuing to pursue academic development. They also used praise to reinforce the positive group processes that contributed to whole-class and group accomplishments and engaged in praise and validation using enthusiastic language. These teacher practices potentially encouraged a sense of academic connection and individual belonging in the classroom.
In contrast to teachers in low/decreasing integration classrooms, those in high/increasing integration classrooms used praise to promote learning in ways that highlighted students’ strengths and development. Regardless of students’ language status, teachers praised them for good guesses, for taking initiative, or to affirm their belief in students’ work or their capacity to complete it. Notably, for EL-classified students specifically, teachers would publicly compliment their language competence inside and outside of class.
Overall, while teachers in both types of classroom praised students for similar reasons and through similar practices, there were differences in how or why teachers praised students in each environment. The praise of teachers in low/decreasing integration classrooms often served to ease students’ academic concerns or convey different expectations for students based on language status. On the other hand, teachers in high/increasing integration classrooms enthusiastically praised students for their academic accomplishments or effort in working with other students to complete tasks, and they highlighted students’ academic improvements and affirmed their abilities, which is an interesting contrast to the differing expectations conveyed to students classified as ELs in low/decreasing integration classrooms.
Connecting With Students Socially
We also examined teacher practices that related to individual social relationships with students, finding a similar number of instances in low/decreasing (153, or 48.4% of the total) and high/increasing integration classrooms (163, or 51.6% of the total) (see Table 3). While the majority of social connections were similar across classrooms, those that were distinct to either low/decreasing or high/increasing integration settings were markedly different in character.
Social connections in both sets of classrooms
Teachers in both sets of classrooms made social connections with students through various uses of humor to make jokes about the academic material, engage in teasing and sarcasm, and at times to disclose personal information about themselves. Teasing was primarily directed toward students classified as non-ELs rather than ELs. 8 They also discussed students’ personal interests, extracurricular activities, home lives, and individual opinions. With EL-classified students specifically, teachers asked about their other classes and their family members, suggesting concern for both their academic and personal lives. In some instances, teachers in both sets of classrooms dismissed students’ offerings of personal connections: in most cases these simply functioned to put boundaries around tangential comments, stories, or questions in order to make room for a focus on immediate academic goals.
Social connections in low/decreasing integration classrooms
The primary practice found to be unique to low/decreasing integration classrooms was teachers’ rejection of students’ requests for materials or movement around the room or of their attempts at classroom participation that appeared to overtly challenge the teacher or disrupt the lesson. For EL-classified students specifically, teachers in low/decreasing integration classrooms were also found to at times reprimand these students for joking or laughing, or in other cases to reprimand students with a joking tone of their own. The potential for misunderstanding in such situations is visible in the following interaction, when Ms. Lane used humor to obliquely reprimand EL-classified students for sharing hall passes: While students are getting out their computers Ms. Lane says something to Alonso about passes (I think he asked if he could go to the bathroom but then said he was out of passes).
9
Ms. Lane then says, loud enough for the class to hear, that she found out a secret which was that Miriam had been giving all of her passes to Alonso so that he had extra. This makes the class giggle. I don’t catch Miriam’s reaction, but when I do look over she is quiet. I can’t tell if she is embarrassed or not.
While Miriam’s reaction in this instance was ambiguous, such uses of humor add a layer of complexity for students who are in the process of learning a new language and culture and may or may not understand the meanings, nuances, and accepted uses of humor in classroom contexts (Duff, 2002).
Social connections in high/increasing integration classrooms
The key practice unique to high/increasing integration classrooms was teachers getting to know students personally and academically. Beyond the discussion topics described above that were identified across both sets of classrooms, teachers in high/increasing integration classrooms also asked about and discussed students’ behavior, personalities, personal styles, personal preferences, prior schooling, and experiences with standardized testing. The following example from Ms. Gibson suggested what this type of involvement typically involved: Luis walks into the classroom. He comes in and says to Ms. Gibson that he just took his WIDA test yesterday. Ms. Gibson asks how he feels about it. Luis says he did OK but writing was “bad” for him. He then grabs the stack of notebooks and places them on each student’s desk before sitting down (it’s interesting that he volunteers himself to do so).
In this instance, Ms. Gibson expressed interest not simply in a student’s academic experience of taking the WIDA test, but rather in the student’s feelings about the exam. The student responded by reflecting on his experience in the exam and then choosing to help Ms. Gibson distribute notebooks, taking on a voluntary task in ways that suggest a level of individual belonging in the classroom.
In sum, teachers in all classrooms exhibited a variety of ways to get to know students and form connections, but teachers in high/increasing integration classroom demonstrated more variation in how they formed these social connections. However, the greatest distinction between low/decreasing integration and high/increasing integration classrooms was that teachers in high/increasing integration classrooms had fewer practices that could cause disconnection. Although containing or bounding student concerns was a practice identified across both sets of classrooms, patterns of shutting students down through rejecting requests and restricting participation, and of using humor with EL-classified students in the context of reprimands, were found exclusively in low/decreasing integration classrooms.
Facilitating Interactive Learning Among Peers
We also coded for how teachers took steps to help peers build relationships with each other, particularly with regard to academic collaboration. To understand the processes involved in teachers’ facilitation of interactive peer work and not simply its presence in the classroom, we coded practices that demonstrated how the teacher engaged with students around peer work in the classroom. We found fewer instances in low/decreasing integration classrooms (60, or 34.5% of the total) than in high/increasing integration classrooms (114, or 65.5% of the total) (see Table 3), and while practices common to all classrooms comprised the majority of all examples we encountered, those that were unique differed notably from each other.
Interactive peer learning in both sets of classrooms
Teachers in both sets of classrooms assigned work in pairs and groups, made sure that students had partners when they were supposed to, and facilitated pair/group tasks by answering student questions or clarifying instructions. In addition, they discussed the importance of working well with assigned (rather than self-selected) partners and praised students for collaborating well with each other. Teachers also assigned both EL- and non-EL-classified students as peer helpers in group work contexts.
Interactive peer learning in low/decreasing integration classrooms
Two features were unique to teacher practices in relation to peer work in low/decreasing integration classrooms: assigning less complex tasks and separating students from each other. Teachers used peer learning structures for simple or less-involved tasks, such as when Ms. Husk asked students to “compare your answers to the answer key” or Ms. Shaw acknowledged during a journal activity that students might “get inspired by what others say” when they shared their answers with each other. In this sense, checking answers and simply encouraging exposure to another person’s work as a way to generate ideas offered relatively restrictive opportunities for peer interaction and engagement in substantial dialogue. Teachers in low/decreasing integration classrooms also frequently offered students the option to work by themselves during pair or group tasks and separated students from each other during peer learning tasks if judged to not be working well together or engaging in off-task conversation.
Interactive peer learning in high/increasing integration classrooms
In high/increasing integration classrooms, peer learning structures were used not only to facilitate content learning, but also as a focus of the learning itself. Teachers often assigned relatively complex peer tasks that required collaboration and used a variety of practices for teaching peer work skills, as appropriate to the given activity or assignment. For example, teachers were observed: discussing at the outset of a task how to work well together (such as how to take constructive criticism from a peer); giving each person in the group a role in a task; checking to ensure that students were working effectively with their partners; and mediating peer conflict occurring during peer learning. For example, both Mr. Hansen and his co-teacher Mr. Wilson intervened to support group processes in the following excerpt: Leah gets up and walks over to Mr. Hansen to tell him, “Andres isn’t letting me download them [videos of advertisements they have created in their groups] and he’s trying to figure it out himself. And I know how to do it.” Mr. Hansen turns to the group and says, “Andres, this is a group thing” before telling Leah, “Just because you know how to do it, you all need to work together.” Leah replies, “Well, he’s not letting me.” Mr. Wilson, who was standing next to Mr. Hansen, asks, “Is it just in your opinion that he doesn’t know how to do this, or is it a fact that he doesn’t know how to do this?” Andres says, “I’m learning.” Leah then says, “Well, I know how to do it and we would get it done quicker.” Leah tries to tell him what to do again, and Andres says, “Calm down!” Leah replies testily, “Fine, I’m not doing it anymore.” While Mr. Hansen is called over by another group, Mr. Wilson joins Leah’s group and speaks calmly to Leah, asking first about a home-related issue before then making eye contact with her and saying calmly, “Everybody’s learning. Everybody’s making mistakes. Be patient.”
After this interaction, although Leah continued to offer to complete the task for Andres, he mirrored the tone of Mr. Wilson’s words, explaining to Leah, “I’m trying, ok? This is my first time using [a computer program] like this.”
Even in the context of less overtly collaborative tasks, teachers also supported interactive peer learning by assigning students classified as ELs as “teachers” to other students, including ELs and non-ELs, when their individual tasks were complete. For example, Ms. Lane engaged Emma, who had already finished her math work, in helping a fellow EL-classified student: Ms. Lane notices that Emma has finished her work and asks Liana if Emma can help her . . . [When she says yes,] Emma goes over, kneels beside Liana, and reads the problem to her using Spanish but inserting the word “times” as the only English word in the problem (the problems are written in English so this is a fairly complex task).
Such patterns—and the spontaneous and bi/multilingual peer support they engender—contrasted notably with the pattern of limited expectations during group work in low/decreasing integration classrooms.
Overall, classrooms with high/increasing integration were distinguished by proactive teaching of interactive peer learning skills through a range of collaborative activities and a broader variety of practices through which teachers facilitated those activities. Both low/decreasing and high/increasing integration classroom teachers included peer learning tasks and made sure every student had a partner, but only in the high/increasing integration classrooms did teachers make sure students were actively working together rather than allowing partner work to devolve into independent individual work. Both low/decreasing and high/increasing integration classrooms also had instances in which teachers emphasized the importance of working with others, but only teachers in high/increasing integration classrooms actually provided scaffolding through which students could develop the necessary skills for doing so.
Providing Behavioral Support
We analyzed teacher practices regarding classroom management through both managing negative behavior and encouraging positive behavior. We found more instances of these practices in low/decreasing integration classrooms (287, or 66.9% of the total) than in high/increasing integration classrooms (142, or 33.2% of the total). Furthermore, the more proactive practices found only in high/increasing contexts were relatively rare, while the more punitive or negative practices unique to low/decreasing classrooms accounted for over one quarter of the total instances overall (see Table 3.)
Behavioral support in both sets of classrooms
Teachers in both sets of classrooms managed student behavior by using a range of redirections, telling students both what to do and what not to do. Nonverbally, they also employed strategies such as clapping or counting to get students’ attention or using hand gestures to silently redirect students. Teachers treated students classified as ELs similarly to non-ELs when they were misbehaving in mixed status (EL and non-EL) groups. When responding to the behavior of some individual EL-classified students, particularly those earlier in the process of developing English, teachers across both sets of classrooms tended to treat them somewhat differently than non-ELs by repeating instructions and prefacing redirections with positive affirmation, perhaps acknowledging that misunderstanding may have caused unexpected behaviors.
Behavioral support in low/decreasing integration classrooms
Classroom management practices unique to low/decreasing integration classrooms included threats of punishment, particularly those designed to exclude students from classroom community or school activities. Such punishments included separating groups of peers, sending a misbehaving student out of the room, or issuing detention or suspension. For example, Ms. Husk threatened to send students out of the room if they misbehaved because they “are seventh graders [and should act like it].” In the context of such antagonism, teachers also engaged in arguments with students in front of other classmates regarding behavior that did or did not occur and the consequences of breaking classroom rules.
Practices were also characterized by a negative tone and antagonistic processes. For example, teachers often called out a student’s name with a negative tone to indicate disapproval but did not provide further redirection for that student. Teachers also used sarcasm to indicate displeasure with student behavior, the connotations of which might have been particularly challenging for some EL-classified students to interpret (Duff, 2002). Teachers would at times state that certain behaviors were annoying to them personally, or “ridiculous” (Ms. Miller) instead of explaining why the behavior would be problematic for the classroom community. In such instances, students may not have understood why their actions were inappropriate for the situation in which they occurred.
Behavioral support in high/increasing integration classrooms
Classroom management practices unique to high/increasing integration classrooms, while relatively rare, were characterized by proactive teaching and coaching of classroom behaviors and the use of a positive or respectful approach to addressing problematic behaviors. One way in which teachers engaged in proactive behavioral coaching was by discussing expectations ahead of a relevant activity. For example, Mr. Brown prefaced student presentations by eliciting from students what the appropriate behaviors would be during that activity: Mr. Brown says that they’ll start presentations now. He talks about his expectations for their conduct during the presentations—what he terms “Looks Like, Feels Like, and Sounds Like.” He asks the class what the students should look like during the presentations. Aaron and Evan raise their hands, and Mr. Brown calls on Evan, who answers, “Paying attention.” Mr. Brown agrees, and talks about what “paying attention” looks like—he says something about keeping their eyes on the projector and their hands free (not doing anything else). Aaron repeats fragments of what Mr. Brown says, saying “Eyes on the projector” and “Hands free, yo.”
Teachers also pointed out positive behaviors to engage other students in doing the same and displayed a positive and respectful approach when correcting student behavior by using phrases like “please” and “thank you,” such as when Ms. Gibson said to two giggling students, “This is an individual task, so please work by yourself.” Dr. King-Porter, for example, also put notes on students’ desk that praised their actions while supporting positive behavior, such as when she wrote to a student who had recently returned from the principal’s office for disciplinary action (unrelated to Dr. King-Porter’s class). Her note read, “You did a great job getting it together when you came back.”
Overall, teachers in high/increasing integration classroom approached classroom management with proactive practices that served to scaffold relevant behavioral skills, communicate respect for students, and build positive rapport. In this way, behavior management strategies employed in high/increasing integration classrooms served to reinforce positive shared norms for both EL- and non-EL-classified students. This was in contrast to sarcastic and potentially insulting or confusing comments found in low/decreasing integration classrooms, where teachers also threatened or used forms of punishment that excluded individuals from the classroom or school community.
Directly Addressing Diversity and Discrimination
The final set of teacher practices were those in which teachers publicly addressed issues of diversity and discrimination. Only two classroom teachers (Dr. King-Porter and Mr. Hansen) did so, and both of them had high/increasing integration classrooms. Dr. King-Porter integrated diversity as a positive classroom characteristic by regularly asking students to form “culturally and linguistically diverse groups” for in-class work and explain to her how their groups were diverse by these criteria. She also emphasized the value of language diversity by explicitly telling students that “people [in class can] speak English and Spanish or English and other languages,” inquiring into the vocabulary or structure of students’ non-English home languages and pointing out her willingness to help students’ parents who do not speak English. Of particular note in relation to issues of diversity and discrimination is that Dr. King-Porter was unique in having completed doctoral-level education (and with a focus on diversity), and she was also the only African American teacher in our study. It is likely that both of these elements played a role in her teaching practices, particularly in relation to issues of cultural and linguistic diversity.
Mr. Hansen—who had not pursued further education on issues of diversity or experienced discrimination as a person of color—also addressed these issues, though he did so using different approaches. In his classroom, Mr. Hansen talked about structural issues related to racism and implicated himself as a White man in society during discussions of the novels that students read as part of the ELA curriculum, using literature as a means to discuss relevant contemporary issues. The example below occurred while reading the novel The Watsons Go to Birmingham (Curtis, 1995), a story set in 1963 about a 10-year-old African American boy and his family during the American Civil Rights movement: Mr. Hansen asks, “It’s a white angel. Why do you think the angel was white?” Elias replies, “Because that’s the way they sell them.” Mr. Hansen picks up on this and says, “Did you all hear that? What he just blurted out? That’s the way they sell them. Has anybody been shopping and walked down the aisles of Toys “R” Us? Anything you noticed about the dolls?”
Mr. Hansen went on to facilitate a further discussion of race and popular culture, with eager participation from students in the class. Although he did not directly discuss classroom diversity in the way that Dr. King-Porter did, he created an opportunity for students to explore the lack of racial diversity in American cultural representations. Given that practices addressing diversity were only found in two classrooms, however, and differed notably from each other, few firm conclusions can be drawn about this set of teacher practices from the data available.
Discussion
Through social network analysis, we assessed classroom linguistic integration of academic peer networks, operationalized as the extent to which peer nominations of academic helping “bridge” across language boundaries between students classified as ELs and non-ELs. Using a dynamic systems framework within the context of ecological theories of language learning, we identified specific teacher practices that were unique to low/decreasing and high/increasing integration classrooms, or shared by both.
Key Findings
Demographic Classroom Characteristics
Teacher practices both influence and are influenced by the characteristics of the classroom, and so a better understanding of the varied ways in which classrooms at distinct levels of segregation may differ from each other demographically is key to realizing the potential of social network tools in understanding linguistically diverse classrooms. In our sample, several structural features which one might assume would influence classroom linguistic integration—such as percentage EL-classified population, average academic achievement (via standardized state assessments), average and range of English language proficiency, student mobility, number of different languages spoken by students in the class, grade level, academic track, and subject—in fact did not. These findings suggest that demographic factors did not appear to be associated with the linguistic integration of classroom peer networks in the classrooms we studied.
Teacher Practices
While we identified many practices shared amongst high/increasing and low/decreasing integration classrooms, we also found patterns of teacher practices distinct to each, suggesting a range of relationships between linguistic integration levels and these practices.
The patterns of praise and validation found in high/increasing integration classrooms were both more frequent than those in low/decreasing integration contexts and more likely to be characterized by specificity, enthusiasm, and positive affect. Teachers highlighted student competence and engagement, particularly for EL-classified students, which aligns with previous work on the ways in which teachers can ameliorate status problems for youth in linguistically heterogeneous classrooms (Bunch et al., 2005; E. G. Cohen & Lotan, 1995; Khisty, 1993). These practices of praise and validation reflected and reinforced a growth mind-set (Dweck, 2006), which has been demonstrated to improve self-efficacy and academic engagement and achievement, though not specifically in studies of linguistically diverse youth. In low/decreasing integration classrooms, in contrast, teacher validation tended to be reactive rather than proactive and to portray a fixed mind-set. While such practices were relatively rare, they—along with patterns of validating effort to mitigate anxiety and conveying differential expectations and rewards for EL-classified students—nonetheless suggest a set of interactions unique to low/decreasing integration classrooms.
Patterns of teacher involvement in developing social relationships in high/increasing integration classrooms revealed that teachers used many different ways of getting to know students, including but not limited to those already identified in the literature on linguistically diverse classrooms (Michael et al., 2007; Pass & Mantero, 2009) as well as some only identified in ESL classrooms (Harklau, 1994; Yoon, 2010). Teachers in high/increasing integration classrooms demonstrated a somewhat wider range of these practices than those in low/decreasing integration classrooms, where instead we found instances of teacher rejection of student requests and bids for participation, in conjunction with uses of humor in the context of reprimanding, which might have further disconnected these students from the classroom cultural context (Duff, 2002).
Regarding peer work, teachers in both sets of classrooms facilitated these contexts for learning, although those in high/increasing integration classrooms were found to do so more often, and to demonstrate proactive teaching of peer work skills through facilitating complex interactive peer learning tasks, positive norms, and open-ended questions—practices that have also been found in existing research on students classified as ELs (Bunch et al., 2005; Khisty, 1993). Unlike the complex problem-solving activities found in studies of successful heterogeneous grouping (Bunch, 2014; Bunch et al., 2005), the simplified tasks often employed in low/decreasing classrooms did not provide the same kind of opportunity for students to engage with each other academically and build peer relationships. In this sense, both the task design and less-frequent facilitation of peer learning activities found in low/decreasing classrooms may have served to limit ecological affordances for language development (van Lier, 2002, 2004) available to develop and demonstrate competence by both giving and receiving peer assistance (Jacob et al., 1996).
In relation to behavioral support, while many common practices were shared among teachers, those that were unique to classrooms at different levels of integration differed notably. Teachers in high/increasing integration classrooms used proactive practices for teaching and coaching students in classroom behavior through positive and respectful approaches, albeit somewhat infrequently. In contrast, teachers in low/decreasing integration classrooms regularly approached students with antagonism and a negative tone. As research has shown, proactive, positive behavior management practices that are put in place to facilitate an environment of teamwork and collaborative effort help students build the skills they need to be members of the classroom community (Bunch et al., 2005).
Finally, two teachers—both in high/increasing integration classrooms— directly and publicly addressed issues of diversity and discrimination. Their practices provide useful insight into ways that both classroom procedures and academic content can be used to directly address concerns raised in García-Coll et al. (1996) about discrimination that minoritized youth may face in schools and other settings. However, given existing scholarship about the challenges of adequately addressing diversity and discrimination in classroom settings (Harklau, 2003; Talmy, 2010), further research is necessary to better understand both the range of practices that teachers in other settings may employ and the impact of these practices on students.
Across these sets of practices, teachers in high/increasing integration classrooms indicated that they valued student individuality by showing respect and recognition for each student as a person. Furthermore, they proactively supported students in skill-building so that they could contribute to the classroom community through managing their own behavior and effort and engaging in productive collaboration with each other. While such patterns were not entirely absent from low/decreasing integration classrooms, the unique ways in which teachers in high/increasing integration classrooms fostered social engagement, facilitated peer work, and encouraged class participation created classroom ecologies in which meaningful opportunities were made available for language development, academic learning, relationship building, and minoritized youth development.
Strengths and Limitations
This study has several particular strengths and limitations. To begin with the former, we are the first to use a peer nomination-based measure of social network integration to quantify EL- and non-EL-classified students integration in linguistically diverse classrooms, and we are also the first to look at social networks of academic helping relationships in this context. This approach, combined with qualitative classroom observations, allowed us to examine classrooms as developmental and ecological contexts from a mixed-methods perspective in ways that both included and extended beyond student and teacher self-report. Furthermore, this study is also strengthened by the dual lens of individual belonging and classroom community for the qualitative analysis, which allowed us to work toward a theoretically grounded, holistic understanding of classroom relationships beyond the relatively more studied phenomena of interactive peer learning. Finally, our extensive qualitative data analysis procedures supported the development of valid, reliable, and detailed findings regarding teacher practices.
This multilayered, multimethod approach also allowed us to benefit from a large sample size for the quantitative social network analysis that was used to select classrooms in the smaller qualitative sample. Therefore, we were able to move beyond subjectively judging teacher practices in each classroom because we could leverage this separate measure of classroom linguistic integration that distinguished low/decreasing integration classrooms from high/increasing integration ones. This differentiation provided an opportunity to triangulate findings and facilitated the iterative analysis of teacher practices, allowing us to identify practices across both sets of classrooms in addition to patterns unique to each group.
Limitations to this study are relevant as well. The findings as yet remain descriptors of dynamic patterns rather than claims to causality. Furthermore, we understand a bidirectional relationship between teacher practices and academic peer networks: Classrooms in which students are more connected to each other, for example, may provide teachers with more support and opportunities for demonstrating positive, socially connective teacher practices. We also acknowledge that the logistical demands of a large mixed-methods study did not allow for additional qualitative observations that might have provided additional detail. Additionally, we were also limited in our ability to member-check results with participants because of the multiyear character of our study. While doing so would have further informed our analysis, the validity checks in place served to mitigate this concern. Finally, we do not claim generalizability of our findings, and acknowledge that our student population is but one of many diverse configurations found in U.S. schools. Our teacher population—while reflective of the dominance of White females in the teaching profession overall—similarly does not address the full diversity of teachers currently serving students across the country. Rather, our study provides a starting point for further investigations of linguistic integration in a range of different classroom settings.
Conclusions, Implications, and Further Research
There is no single linguistically diverse classroom, and both EL- and non-EL-classified students must be understood in their full complexity and variation, rather than as members of monolithic groups. Classrooms must also be conceptualized as nuanced ecological spaces, ones in which teacher practices—as they relate to both students’ sense of individual belonging and the communities students develop among peers in the classroom—have the potential to create supportive rather than discriminatory developmental contexts for minoritized youth (García-Coll et al., 1996), as well as to provide settings in which teachers and peers are ecological affordances for language development (Kramsch, 2002). The linguistic integration of academic peer networks offers a theoretically grounded approach to explore these developmental settings at multiple levels, though by no means is it the only avenue through which such dynamics can be examined.
This study contributes to such efforts by providing an empirical foundation for mapping out demographic characteristics and teacher practices associated with differences in the linguistic integration of students’ academic peer networks. While many aspects of the teacher practices we found align with other more general models of effective teaching, those described in this article are unique in their explicit attention to EL-classified students and diversity more generally. They also illuminate the ways in which classroom linguistic integration is likely related to a multifaceted set of teacher practices rather than singular strategies or actions.
There is no direct prescription for what teachers can do to foster connection and collaboration among adolescent EL- and non-EL-classified students, and our findings are not intended to prescribe a set of practices for specific use in linguistically diverse classroom settings. They instead provide for several layers of application. At one level, they for allow for a deeper scholarly understanding of the complexity of building academic communities across language differences and the varied ways in which teacher practices may play a role in the development and maintenance of integrated classroom communities for adolescents. This study also demonstrates specific directions to pursue in supporting teachers’ efforts to foster social connections among students of varying language statuses in ways that may support linguistically integrated classroom communities. It is important to note, however, that individual teacher practices do not exist in isolation, and any approach to the development of linguistically integrated classroom communities must be implemented holistically, with a clear understanding of students’ instructional strengths and needs and the unique composition of any classroom.
Both the methodology and the key findings of the present study can inform additional research on how teachers might contribute to classroom integration. For example, future research needs to explore how our findings compare with those in settings that differ from this one geographically, linguistically, and by ethnicity/race, age, and other student and teacher demographic factors. Studies can also follow individual teachers across multiple classes and years to better understand the multidirectional relationships among linguistic integration and teacher practices, as well as the ways in which these classroom ecologies are situated in and influenced by varied and complex sociopolitical contexts. Future research can additionally address students’ perceptions of particular teacher practices and their impacts, as well as teachers’ reflections on their implementation.
Footnotes
Notes
A
L
V
M
R
H