New Teachers Learning Disciplined Improvisation for Meaningful Talk in Diverse Classrooms

Conceptions of Discussion, Teaching, Teacher Learning

We clarify conceptions. By discussion, we mean group structures for recall, language production/elaboration, exploring interpretations, linking content and community, and working through problems and conflicting ideas. Teaching through discussion includes substantive engagement associated with academic learning, using authentic questions, uptake for cohesive discourse, and high-quality feedback (Applebee, Langer, Nystrand, & Gamoran, 2003; Nystrand, 1997). Such discussion facilitates discovery and display through language (Bunch, 2013); scaffolding language of schooling (Schleppegrell, 2004); and leveraging linguistic resources in students’ communicative repertoires (Martinez, Morales, & Aldana, 2017). Such teaching includes disciplined improvisation with moments co-constructed in real time. By disciplined we refer to content purposes for discussion, repertoires of knowledge and techniques (Erickson, 2011), repertoires of practice (Gutierrez & Rogoff, 2003), structures and constraints (Hammerness et al., 2005; Sawyer, 2004). By improvisation we mean in-the-moment decision-making and microadaptations (Barker, 2012; Corno, 2008) to focus engagement and learning.

Relatively little is known about how teachers learn to enact discussion: conceptual and practical tools needed (Williamson, 2013), adaptations teachers learn to make, a typology of moves available. A slim literature finds microadaptations in literacy lessons include: modifying objectives, changing means by which objectives are met (adapt routines, strategies, materials, actions), inventing examples or analogies, inserting mini-lessons, suggesting different ways students can handle a problem, and resequencing or omitting activities (Duffy et al., 2008; Parsons, 2012). Adaptations require keen perception, and issues warranting adaptation may go unnoticed, especially by novices who may not see “telltale signs for adjustment” (Duffy et al., 2009, p. 168). Pattern recognition requires development, as it is seldom a natural process for teachers (Korthagen, 2010). Challenges of facilitating oral language activity intensify for many new teachers underprepared to support learners in linguistically complex classrooms (Ball, 2009; Darling-Hammond, 2006; Gándara, Maxwell-Jolly, & Driscoll, 2005), such as those we feature in our project.

As classroom talk may be fast-paced and ephemeral (Wortham, 2008), adaptation in-the-moment is demanding. Teacher development that supports perceptual skills in noticing language contributions, nonverbal cues, and efforts at discourse, and metacognitive processing of action and talk may support teachers’ capacity to reflect-in-action (Schon, 1983; Yanow & Tsoukas, 2009) and make informed choices that support learning. We need to know much more about adaptations a teacher might make and ways teachers learn to enact them.

Cognition and Teacher Learning for Discussion

Our project investigates cognitive underpinnings of disciplined improvisation, a form of adaptive expertise (Bransford et al., 2005) that benefits from conceptually rich knowledge of teaching practices and their effects. Like other expert decision-making, disciplined improvisation involves knowledge-driven and perceptual-driven processes (also called top-down and bottom-up processing; Chi, Feltovich, & Glaser, 1981). For example, teachers must foster discourse incorporating all voices (e.g., noticing potentially relevant observations voiced outside of “standard” academic language). A number of studies explore teacher noticing (Sherin, Jacobs, & Phillip, 2011), but perceptual aspects of disciplined improvisation have received little attention.

Instruction focused exclusively on building repertoires of knowledge typically fails to influence practice (Bransford et al., 2005). This problem of enactment arises from a paucity of opportunity to rehearse these learned strategies. Lack of focus on perceptual processing likely exacerbates the problem, as teachers may fail to notice conditions of applicability. We will develop both a process model (how does disciplinary improvisation work in-the-moment) and a learning model (how do teachers learn to enact disciplined improvisation) to support classroom discourse in linguistically diverse classrooms.

There remains little empirical elaboration of underlying processes of disciplined improvisation. Sawyer (2011) argues that teaching expertise is “mastery of a corpus of knowledge—ready-made, known solutions to standard problems–but in a special way that supports improvisational practice.” In our simplified process model (Figure 1), disciplined improvisation is informed by knowledge and practice, informed by perceptual learning of meaningful patterns, and ultimately responsive in-the-moment—through microadaptations—to diverse students’ learning, language, and lives. Metacognition—awareness of one’s own knowledge and skill, and ability to control one’s cognition—knits these pieces together. Our research will elucidate repertoires of knowledge and practice and perceptual skills—and relationships between them—necessary to support discourse in linguistically diverse classrooms. 

Our initial model of how disciplined improvisation can be learned (Figure 2) is also highly simplified: teachers need to develop their repertoires of knowledge and practice, they need to develop perceptual skills, and they need ample opportunities to develop the metacognitive skills needed to integrate knowledge-driven and perception-driven processing in authentic contexts. Our definition of teacher learning is growth in any of these three areas. With teacher partners, we will design educational experiences for early-career teachers so we can observe, analyze, and improve designs for teacher learning across these three areas.

Refined Research Questions

Our overarching question is: 

How do teachers learn to enact disciplined improvisation in support of discourse in linguistically diverse classrooms?

We examine this overarching question through three main sub-questions. For each, we look at the nature of the expertise (process model) and how it is learned through experience (learning model):

RQ1. Learning the dialogic toolkit

• What knowledge and practices support disciplined improvisation in linguistically diverse classrooms?
• What learning activities and tools support their development?

RQ2. Learning to see in discourse

• What perceptual/noticing skills support disciplined improvisation in linguistically diverse classrooms?
• What learning activities and tools support their development?

RQ3. Enacting disciplined improvisation in the moment 

• What metacognitive skills support the integration of knowledge, practice, and perception into disciplined improvisation?
• What learning activities and tools support their development?

Alignment with Program Goals and Response to Feedback

Early-career teacher learning. We have found even when preservice teachers (PSTs) develop rich repertoires of knowledge and practice for discussion through coursework, first enactments are fraught with complexity, needing greater retrieval of coursework information and knowledge synthesis. TE needs innovative pedagogical models to develop agentive educators and engage teachers in research-informed practice, particularly in light of “reforms” stripping TE of its research-based and visionary roles (Zeichner, 2014). As part of a McDonnell-funded collective, we will add focus to preservice/early-career teachers and models of their learning and of TE pedagogy to support development of novices for communicative work in diverse classrooms. 

ELA, then science. Responding to feedback, we examined ways our design-based research with TE and teacher partners has yielded rich extant databases in ELA that inform study plans. Emergent findings will inform focus on elementary science, so we can test our models, aligned with the program call for “cross-cutting impact across subject areas and ages.”

Scale-down. We focus now on nuanced examination of teacher learning and model development in California’s richly diverse population, eliminating the UMD bilingual teachers team. We also no longer will study dissemination with CSU, instead tapping CSU faculty as advisors and going deep with preservice and early-career teachers in local schools.

Virtual reality (VR). Feedback raised VR concerns: transferability, “authenticity” versus live improvisation, cost/sustainability. We recognize concerns but find VR a promising, rapidly-evolving avenue with affordances that can be leveraged, combined with face-to-face teacher learning supports (teacher educator feedback, professional learning communities). Opportunity for iterative change over time in a lower-risk environment (real students not impacted as teachers practice) is a significant VR affordance, as is creation of tailored, reusable learning environments ensuring PSTs sustain practice with important discussion aspects to which they may have uneven exposure in real-world settings (e.g., facilitating discussion in linguistically diverse classrooms, distinguishing uptake/pseudo-uptake). 

Another key VR affordance is ability to record authentic interactions to return to and iterate upon, making ephemeral discussion a durable artifact for analysis and improvement. When people interact with human-controlled avatars versus algorithm-controlled agents, their physiological reactions and behaviors are similar to how they would react to real people (e.g., Donath, 2007; Hoyt, Blascovich, & Swinth, 2003; Okita, Bailenson, & Schwartz, 2008). This suggests that immersive virtual environments with avatars that we would use may provide novel, contextualized classroom talk situations that are realistic, potentially augmenting face-to-face preparation by repeatedly drawing teachers’ attention to consequential elements of classroom discussion that can go unseen or unexplored in real-world settings. 

We are aware that VR study reviews reveal uneven results (e.g., Khan et al., 2009; Laver et al., 2015), and there is need for rigorous research that explicitly addresses transfer from virtual to real-world settings. However, we are encouraged by some investigations suggesting cognitive strategy elements of these two environments are broadly equivalent and that there is a clear positive transfer effect from virtual to real training when virtual task elements closely approximate real-world task elements (e.g., Rose et al., 1998, 2000). We will cautiously build on this promising research by learning about teachers’ local contexts and histories in classroom discussion practices, then designing virtual environments that reflect realities of those contexts. Additionally, our approach to transfer differs from many reviewed studies (particularly in rehabilitation) which feature spontaneous transfer. We will support teachers in guided transfer through activities designed to bridge learning between virtual and real-classroom environments. Finally, we intend to contribute to literature on VR for training with systematic reporting of study details and investment in understanding teachers’ learning and development across virtual and face-to-face settings over time.