Investigating discourses for administrators’ learning within instructional rounds

Introduction

Calls for instructional leadership at all levels of school administration are pervasive across the educational landscape (Brazer and Bauer, 2013; Hallinger, 2005; Honig, 2012; Neumerski, 2013; Noble et al., 2010). While variously defined and enacted, the overarching conception of instructional leadership is that of school leaders who put improving classroom practice at the center of their work, in contrast to other conceptions of leadership that might focus school leaders’ attention on organizational management or democratic decision making. However, while the term ‘instructional leadership’ has gained in popularity since it developed from the effective schools literature (Edmonds, 1979; Salo et al., 2014), the primary conclusions highlight the importance of a strong instructional leader rather than set practices that instructional leaders engage in to make their schools more effective (Neumerski, 2013).

In addition to the rhetorical use of the term, other models for instructional leadership outline specific leadership practices (Leithwood et al., 2004). Of particular note is Hallinger’s model (Hallinger, 2005; Hallinger and Murphy, 1985), based on the idea that the capacity for effective instructional leadership relies upon leaders’ knowledge of the instructional core, that is, the interaction among student learning, instruction, and content. He recognizes three dimensions of instructional leadership: (1) defining the school’s mission; (2) managing the instructional program; and (3) promoting a positive learning climate. For each, he identifies specific accompanying practices in which instructional leaders engage. Addressing the second dimension around managing the instructional program, Hallinger noted three specific practices: (1) supervising and evaluating instruction; (2) coordinating the curriculum; and (3) monitoring student progress; all involve school leaders’ engagement in day-to-day instructional practices.

While good leadership requires more than a leader carrying out a given set of practices, the research on how school leaders actually enact these practices remains underexplored (Neumerski, 2013). One specific professional learning practice that is posited to support leaders’ capacity is instructional rounds. In instructional rounds protocols, teams of school and central office administrators (and/or teachers) observe classrooms and document their observations in relation to instructional issues, referred to as ‘problems of practice’, posited by the school or district, and then provide feedback to the host school (City et al., 2009; Fowler-Finn, 2013; Roberts, 2012). Instructional rounds is grounded in collective inquiry: through the collective analysis of evidence from classroom observations, school leaders, at both the school- and district-level, have the opportunity to develop and enact their understandings of the instructional core (City, 2011).

The practice of rounds also lends itself to the reconceptualization of instructional leadership as ‘leadership for learning’—in which the focus is not so much on what teachers are doing, but more so on student learning (Bush, 2014): the classroom observations central to the instructional rounds process supports an emphasis on what students are doing. At the same time, the collective nature of rounds aligns the practice with more recent conceptions of transformative and distributed leadership (Marks and Printy, 2003; Salo et al., 2014).

In this article, we examine instructional rounds visits conducted by a network of school district superintendents in a northeastern state of the USA over the course of five years. In particular, we investigate how whole-group discussions afford and constrain opportunities for participants to deepen their understanding of the instructional core, a foundation for instructional leadership practices, especially managing instruction.

Central to instructional rounds as a learning practice, grounded in inquiry, are the conversations (in small groups or teams and the whole group) about data (or ‘evidence’) from classroom observations. For this reason, our analysis focused on recurring conversational routines within the network’s instructional rounds visits. In analyzing these routines, we employ an analytical framework for discourse we identify as instruction-specific talk (IST). IST describes discourse that focuses on observed classroom episodes within which participants employ specific evidence from observed episodes in constructing knowledge about instruction and student learning. Our analysis addresses three related questions.

When does IST occur within instructional rounds visits?

We conclude with implications for network-level practices focused in developing school- and district-level administrators’ actionable knowledge of the instructional core, as well as implications for further research.

Conceptual Framework

Instructional rounds, fundamentally, is a professional learning model (City, 2011). It is premised on conceptualizations that equate learning with participation in social activity (Rogoff et al., 1995; Wenger, 1998; Wertsch, 1991). Learning, in this view, is mediated through the use of specific tools and symbol systems, especially language (Vygotsky, 1978). For this reason, attention to discourse and discourse structures is critical to understanding how social activity contributes to learning, of individuals and groups.

The presence of discourse on its own does not guarantee learning. Instead, discourse structures may be said to afford or constrain opportunities for learning (Little, 2003). In seeking to understand how classroom discourse (teachers’ and students’) afford or constrain opportunities for learning, for example, researchers contrast two types of discourse patterns. One pattern models ‘IRE/F’, in which the teacher Initiates discussion; the student (or students) Respond(s); and the teacher Evaluates and/or offers Feedback. In contrast are those in which teachers invite explanations and probe students to expand their thinking, and students listen to, refer to, and even disagree with one another’s comments (Cazden, 1988, 2001; Lampert, 1990).

Discourse structures that promote the latter include ‘exploratory talk’ (Barnes, 2008; Barnes and Todd, 1977) and ‘accountable talk’ (Michaels et al, 2008). In these discourse frameworks, the emphasis is on constructing new knowledge (ideas, hypotheses, interpretations) through discussion that encourages participants to offer tentative ideas and to respectfully challenge the ideas of others. Explicit in these frameworks is the necessity to support claims with specific information or evidence:

Accountability to knowledge means that when speakers make an observation or claim to support, they try to be as specific and accurate as possible, always maintaining attention to evidence, truth, and the community’s accumulated knowledge. If challenged, they offer evidence that meets the standards of the discipline under discussion. (Michaels et al, 2010: 30)

Frameworks for discourse, including exploratory talk and accountable talk, with their emphasis on using evidence, striving for specificity, and building upon and challenging other participants’ observations contribute to the instruction-specific talk framework used to investigate discourse in this article.

Episodes of discourse available for analysis are the ‘conversational routines’ that recur within a group’s ongoing activity, for example, instructional rounds visits. Following Feldman and Pentland’s (2003: 95) analytical category of ‘organizational routines’, that is, ‘a repetitive, recognizable pattern of interdependent actions, involving multiple actors’, Horn and Little (2009: 184) posit ‘conversation routines’ as ‘patterned and recurrent ways that conversations unfold within a social group. Routines are constituted by “moves,” or turns of talk that shape the interaction’s progress by setting up and constraining the response of the subsequent speakers’. In their study of two discipline-based teacher groups in a high school, these authors identified routines that manifested ‘rough draft talk’, questioning that invites elaboration, and ‘linking moves’ that create a sense of affiliation and membership.

A critical element of conversational routines that afford participants learning opportunities is the relation of more generalized theorizing and questioning with the use of specific details related to the object of study (for example, classroom instruction). Horn (2010: 197) describes a ‘movement between the particular and the general…as a means of developing teaching knowledge that is deeply rooted in embodied accounts of classroom life, joining important concepts about teaching to particulars of practice’. ‘Particulars of practice’ might take many forms, for example, a teacher’s ‘replay’ (Horn, 2010) of an episode of classroom interaction, a direct reference to a sample of student work, or a quantification of classroom activity (for example, time spent on a specific task, number of students speaking during a discussion, and so on).

Similarly, in a study of collective creation within middle school and high school teacher inquiry groups, Allen (2013) demonstrated that groups that were more successful in developing resources for analyzing instruction and supporting instructional improvements: (a) identified specific forms of evidence (from student work, classroom observations, and so on); and (b) engaged in more extended discussions of the evidence. Groups that generated fewer or weaker resources were more likely to offer general evaluations from the same data sources and move to conclusions or recommendations earlier.

It follows that discourse structures that afford opportunities for professional learning, for instance, about the instructional core, are those that model robust and extended use of specific evidence. Indeed, opportunities for learning are threatened when conversations attend more to following protocols than the substance of the inquiry, that is, ‘data as source of evidence and inquiry’ (Timperley and Earl, 2009).

Building upon the discourse frameworks cited above, we employ the analytical category of IST, that is, episodes of discourse that are characterized by (a) use of specific evidence in constructing, proposing, and challenging claims about instruction, and (b) the density of how specific evidence is used, that is, how often it occurs within a given episode of conversation. The presence of these features, of course, does not guarantee professional learning; however, it affords opportunities for learning that, absent these features, are not available.

Methodology

This qualitative study draws on five years of data from a network of superintendents engaged in the practice of instructional rounds to better understand how this practice supports instruction-specific talk and thus supports their instructional leadership in terms of their understanding of the instructional core, a prerequisite for managing instruction.

Data Analysis

Data analysis employed a ‘top-down’ methodology, which focuses on how meaning is constructed within larger conversational units, rather than in discrete turns (Erickson, 2004). We began by looking across all transcripts to identify the types of conversations that the Network engaged in, which we refer to as conversational routines. We then had two researchers code the same transcript using concepts from Michaels et al. (2008) work on accountable talk, including ‘building and extending’, ‘striving for specificity’ and ‘using evidence’. However, in so doing, we realized that focusing on moves diverted us from our focus on instruction, so instead we used Erikson’s top-down coding to identify sustained episodes of IST. We defined sustained episodes of IST as sections of transcripts that included comments about specific aspects of teaching and learning observed during the classroom observations in which at least two people participated for at least two turns each.

Only whole-group discussions were audio recorded and transcribed; thus, small-group discussions were not analyzed. Because of the number and length of recorded whole-group discussions in the data set, we developed a random sampling technique to select conversations from across the Network’s first five years to analyze for sustained episodes of IST. Multiple conversational routines from each instructional rounds visit were included within the sample. For the ‘group processing of findings’, for example, a routine which only occurred in the first two years and in the fifth year, we began by selecting the first meeting, and then choosing every other meeting in which this routine occurred. Within each episode of IST, we looked for the type of comment or question initiated the episode, what, if any, distractors occurred, and how the episode ended. Finally, we employed visual data displays (Miles and Huberman, 1994) to identify patterns across meetings for when, and under which conditions, discourse modeled instruction-specific talk.

OPEN IN VIEWER

Figure 1.

Instructional rounds visit sample agenda.

Findings

In the following sections we report findings from the data analysis described above. The first section responds to the first research questions: (a) When does IST occur within instructional rounds visits? The second responds to research questions (b) What are the conditions that initiate IST? and (c) What are the conditions that threaten IST?

Conversational routines in rounds visits

We identified four types of conversational routines across the 16 rounds visits conducted by the Network in its first five years: conversations with the host, share outs, whole-group discussions, and reflections. Thirteen of the rounds visits included one of each of the four types of routines. Two visits did not include conversations with the host school or district, and one visit did not include a reflection. These variations depended on the host superintendent’s interests and school schedule considerations; however, for the most part, rounds visits followed a similar pattern with similar elements.

Conversations with Hosts

All of the instructional rounds visits began with a presentation by members of the host district’s and/or school’s leadership team, often including the superintendent, instructional supervisors, principals and/or assistant superintendents. These presentations included data about the problem of practice that would be the focus of the observations, for example, test scores, staffing reassignments, curriculum or school-wide initiatives (for example, ‘Let Me Learn’, a curriculum based on identifying student learning styles). At 11 of the 16 meetings, network superintendents had the opportunity to ask questions of clarification after the presentation. On average, these discussions lasted 10.5 minutes, and consisted of superintendents asking specific contextual questions, for example, ‘Is percentage of Special Ed. consistent with district level at this school?’ Typically, the hosts responded to the question briefly. There was no extended conversation in response to questions, and, thus, no possibility for IST.

In addition to the pre-visit conversations, in 13 meetings, at the end of the rounds visits after the superintendents presented their findings, questions and/or suggestions for next steps to the host district or school team, there was a conversation between the superintendents and the host teams. These conversations lasted, on average, about 13 minutes, and often consisted of host team members asking clarifying questions of the superintendents’ presentations or asking for feedback that was not related to the problem of practice. For example, in one of these post-rounds conversations, a teacher whose classroom had been visited asked, ‘Can you address strengths you witnessed across the board?’ Hosts also used this time to provide details about other initiatives in the district. Similar to the pre-visit conversations, for the most part, these conversations involved limited dialogue between questions, thus little or no possibility for IST.

Share Outs

This conversational routine occurred after the superintendents had conducted their classroom visits in small groups. Each small group then met and discussed their observations, with facilitation from one of the network’s design team members. When the whole group came back together, each small group presented their findings, questions, or next steps. Generally, one person from each group read a poster that had the group’s notes written down. After the share out, participants from other groups asked clarifying questions. These questions were intended to be for clarification, not for greater discussion, and for the most part, involved a participant asking about the specific data behind a reported finding, for example: ‘When you say that you saw evidence that all of the students were engaged, how did you determine that?’ In these cases, someone from the group sharing out would then describe the data that they drew on to come to that conclusion, and there was no further conversation, thus no possibility for IST.

Whole-group Discussions

Whole-group discussions took place after the small-group share-outs. In the network’s early years, these discussions were preceded by a ‘dots’ activity. In this activity, the chart paper with findings and ‘wonderings’ (probing questions) from each small group were placed around the room; superintendents placed stickers (‘dots’) next to those findings they deemed most relevant to the host district’s problem of practice. After placing the ‘dots’, the network engaged in a whole-group discussion in which the superintendents attempted to come to consensus on three to five findings across all of the small groups to share with the host team. These discussions began with the findings that had received the most ‘dots.’

Over time it became apparent to the Network that in synthesizing findings from individual groups and presenting to an audience from the host district, the agreed upon findings lacked specificity to become acceptable to all and/or were ‘watered down’ to be palatable to the host school. During the third year, the network eliminated the use of dots and the requirement that the whole group agree on a set of findings. Instead, the host team was invited to be present during the small-group share outs. These share outs were presented as data the school and district staff might use for its own analysis and formulation of next steps.

Starting in January of the network’s fifth year, the network began to discuss the small group share-outs as a whole group again; it is not clear from the transcripts what initially motivated this change, however, participants’ reflections suggested that they valued this approach. In January, the whole group discussed their questions and ideas for next steps, and in the April and May visits, the whole group discussed findings as well. The conversations that focused on discussing findings were on average 36 minutes long, and those that focused solely on ideas for next steps were on average 15 minutes long.

The whole-group conversations generally began with a design team member asking for a superintendent to recommend a pattern for consensus. The design team member might also ask for confirming or disconfirming evidence of that pattern. In these conversations, there were several instances of IST, in which superintendents discussed issues such as cognitive demand or student engagement. In addition, the network often discussed the process of instructional rounds, including what could be determined during a 10–12 minute visit and what they should be focusing on in the conversation, based on the problem of practice.

Reflections

All of the rounds visits, except for one, concluded with a reflection discussion on the visit as a whole, often employing a ‘plus’ (what worked well)/’delta’ (what could be improved) protocol. The majority of the conversation during reflections used this process. For example, one superintendent reflected that ‘something I enjoyed today…was dialogue with the teachers’ during the post-visit discussion with the host school. Superintendents also commented about the effectiveness with which the group had enacted the instructional rounds model. In this regard, one noted, ‘I thought it was deeper conversation on instruction than we’ve had in the past’. The majority of ‘deltas’ also focused on the rounds process, including scheduling, assigning roles for classroom observations, and identifying a strong problem of practice, rather than focusing on specific evidence from classroom observations that might initiate IST.

Where Is Instruction-specific Talk in the Conversational Routines?

IST occurred only in the whole-group processing conversations, as shown in Table 1. While participants did make instruction-specific comments in the other routines, particularly the share-outs of findings, these comments were not taken up by the group and did not lead to sustained conversations (more than two speech turns) between participants around instruction. In the next section, we consider the conditions for instruction-specific talk in whole-group discussions.

OPEN IN VIEWER

Table 1.

IST in conversational routines.

Routines sampled	Number of routines in the sample	Number of routines with IST present
Share outs	6	0
Whole-group discussion	7	5
Conversations with host school/district	6	0
Reflections	6	0

Conditions for Instruction-specific Talk

IST does occur in whole-group discussions in which participants process the findings of the small classroom-visit groups. The transcript excerpt from an early instructional rounds visit (Table 2) provides an example of IST from a whole-group processing discussion in which participants offer specific evidence from classroom observations, in this case, in response to a problem of practice related to higher order thinking skills. As this was part of an early rounds visit, the Network’s aim is to arrive at consensus around three to five findings. In the sections below, we analyze the effect on IST of specific conversational turns.

OPEN IN VIEWER

Table 2.

Transcript excerpt from a whole-group processing discussion.

Facilitator and participant dialogue	Effect on IST
Facilitator 1 [summarizing from the frequency of dots next to individual findings on the chart papers]: ‘Limited evidence of upper level of Bloom’s taxonomy, which may look like lower level tasks, mostly tasks at remembering level, across 5 classes [visited]. That is a persistent issue here…Evidence of frameworks, routines, structures, such as Do Now, objective on the board, and other things…Anything else? Those seem to be the two.	Summarizing themes identified by small groups
Facilitator 2: Should we probe? Make sure there’s no disconfirming evidence or qualifications on those before we decide to put them forward?	Asking for evidence
Facilitator 1: How about the [finding] on lower-ordered thinking skills, the types of tasks and activities we saw in classes? We ought to talk about where we saw something else happen. Group 1, we reported out what we saw in the science class, was clearly higher-ordered, middle and lower, depending on students’ needs. Other examples?	Asking for evidence
Participant 1: In social studies, we saw two or three students push out challenging, higher-order kinds of questions.	Sharing evidence
Participant 2: Do you have an example of one of those questions?	Asking for evidence
Participant 3: Let me find it. About taxes.
A cell phone buzzes, and discussion of norms ensues (4 speaking turns)	Interruption
Facilitator 1: Other evidence?	Asking for evidence
Participant 4: The problem is our own competency with Bloom’s taxonomy. We had a list and verbs, but no exemplars. We all bring in our own experience and biases to this. At one point, I saw applying, but someone else might see it was remembering. We probably didn’t get further than applying, but we did see artifacts.	Discussion of process
Facilitator 1: Do you have an example of those artifacts?	Asking for evidence
Participant 4: That science class. They were showing educational videos of black holes. There were rockets on the side with some writing.	Sharing evidence
Participant 5: The [finding] about higher-order thinking and relevance, in my experience today, did come from archival data [posted student work], not the active teaching and learning process in those classes today but in the evidence of student work that had been previously done…So when we talk about, is it evidence of higher level thinking in Bloom’s terms, it may be, but it may not be, depending upon how that artifact was derived.	Relating evidence to POP
Participant 6: People are emphasizing the narrative that went with the [science] project. If the learning specification or expectation was creative writing, we might have something that is an application. If the learning expectation was that students will understand something about space travel or rocketry, there was no evidence of application there. [However,] there was evidence of creating understanding in this one science class where they were dissecting a squid. The way the questions were asked, students had to draw on past knowledge and come up with a solution, such as the gender question — what gender is the squid?	Relating evidence to POP
Participant 5: Related to what [Participant 6] was saying about the writing, creative or writing about science, I raise the question about how the artifact was created, because in some instances in those papers, there was a series of words underlined. It suggested to me that there were key terms that had to be embedded [by the teacher] in the writing assignment, such as ‘universe,’ ‘telescope,’ ‘observatory,’ as if to say, here are some terms we’ve learned. Now embed them in the story you write.	Using evidence to frame question of interpretation
Facilitator 1: I think the best we can hope for is that we’ve been specific enough [so] that [the host superintendent] and the staff can make sense of the parts. I don’t know that we can do more than that. I think they know which teachers we’re talking about. Those pieces of whether it’s creativity or creation of understanding, that’s something they can follow up on and take a look at. I think we, as a network, ought to struggle with that to give better feedback, all of us.	Process comment

We identified 20 episodes of IST within the seven whole-group processing discussions in our sample, including the one presented in Table 2. Some processing discussions had multiple episodes (as many as eight); most had one or two; and two had none. We analyzed the episodes in relation to these guiding questions: (1) What initiates IST? (2) What threatens IST? (3) How is IST re-engaged when it has been interrupted?

What Initiates IST?

In eight of the 20 episodes in our sample, IST is initiated by a request for evidence from classroom observations (five by a facilitator; three by a superintendent participant). Other initiators are: questions related to the instructional core (for example, about an instructional task) (5); proffers of evidence from a classroom observation (4); or disagreements on the interpretation of evidence presented from classroom observations (3).

In the sample transcript excerpt (Table 2), IST is most evident in the exchange among Participants 4, 5, and 6 that occurs in the latter part of the episode. Here we see reference to specific examples from classroom observations, including student-constructed products (‘rockets’), samples of student writing posted on a classroom wall, and questions asked by students during an activity (dissecting a squid). The participants relate these to a guiding problem of practice (about higher-order thinking) and use them in questioning or challenging one another’s interpretations (for example, Participant 5: ‘Related to what [Participant 6] was saying about the writing, I raise the question of how the artifact was created …’).

IST here is initiated by repeated requests for specific evidence made by the two facilitators (and, in one case, by a participant). For instance, after Facilitator 1 summarizes themes identified by small groups on chart paper, Facilitator 2 probes for ‘disconfirming evidence’, that is, evidence that might contradict the identified themes. After a participant shares an observation about students ‘pushing out challenging, higher-order kinds of questions’, another participant asks for ‘specific examples of those questions’.

Eventually, and despite interruptions (discussed below), these requests for specific evidence encourage the episode of IST identified above.

What Threatens IST?

In 8 of the 20 episodes, IST was temporarily interrupted but subsequently re-engaged. In all cases, it is ultimately curtailed. Threats to IST take these forms: comments about the process of rounds itself (9 instances); synthesizing or generalizing comments (for example, ‘I think we call can agree we saw scaffolding …’) (8); ‘wonderings’ (probing questions) (4); off-topic comments (including jokes) (4); or recognition of inability to agree on a finding (3).

In the transcript extract, there are two brief interruptions to IST, both related to process. The first occurs when a cell phone buzzes and a discussion ensues about the network’s norm about using electronic devices during visits. The second occurs soon thereafter when a participant suggests that the group’s ability to identify evidence related to the problem of practice may be limited by its ‘competency with Bloom’s taxonomy’. As we describe below, neither process discussion curtailed IST; it is re-engaged by a facilitator’s asking for specific evidence from observations.

The episode of IST is finally curtailed by a summary comment about the process. Facilitator 1 comments on the group’s struggle with enacting the instructional rounds practice: ‘You can see that we’re struggling with this process. We’re a learning community’. As if speaking for the network, he views the practice as limited in addressing the identified problem of practice in a way that will be useful to the host school and superintendent.

How Is IST Re-engaged?

As noted above, not all threats to IST are fatal. In some cases, IST may be interrupted temporarily and then re-engaged. In eight instances in our sample, IST is interrupted and re-engaged. The causes for interruption are largely the same as curtailment: off-topic comments (4); process comments (2); and synthesizing comments (2). In five of these instances, IST is re-engaged by a facilitator or participant asking for evidence in relation to the (interrupted) theme or topic. In the remaining three, the facilitator restates an identified pattern or makes an observation about a theme related to the (pre-interruption) theme or topic.

In the transcript extract above, a beginning discussion of evidence related to students’ use of ‘higher-order…questions’ is first interrupted by a buzzing cellphone and subsequent brief discussion of the network norms. It is re-engaged by Facilitator 1’s question, ‘Other evidence?’ and Facilitator 2’s request that a participant share ‘evidence of what might be higher order [thinking]’ from artifacts observed by their small group. IST does not immediately resume at this point. Initially, the superintendent does not provide specific examples; instead, he makes a process comment about the group’s ‘competency’ with applying Bloom’s taxonomy (as the host superintendent’s problem of practice called for); however, when prompted again for an example, he does provide one (a video from a science class). This initiates the exchange identified earlier among Participants 4, 5 and 6.

Discussion

Instructional rounds is an inquiry-based professional learning practice intended to develop participants’ knowledge of the instructional core and thus their capacity to enact instructional leadership (Hallinger, 2005). Participants’ learning in this practice is posited to occur through discourse. For this reason, it is critical to understand how specific discourse structures, that is, the conversational routines we identified within the overall instructional rounds protocol, afford or constrain opportunities for the kinds of discourse thought to promote learning.

Conversational routines are readily identified within instructional rounds visits. Not surprisingly, those routines that deliberately involve participants in discussing findings derived from classroom observations are the ones that manifest the only instances of IST, especially the whole-group deliberation on findings to report to the host school and superintendent. These discussions demonstrated the greatest capacity for ‘movement between the particular and general’ that is understood to develop knowledge of the instructional core (Horn, 2010). They embody a norm of ‘accountability to knowledge’, in which participants, ‘try to be as specific and accurate as possible’ when making an observation or claim (Michaels et al., 2010: 30).

As our analysis demonstrates, and the transcript excerpt (Table 2) illustrates, IST is possible given certain conditions. Particularly generative of IST is the repeated press for ‘evidence’ (examples or artifacts from the classroom observations) and for disconfirming evidence by facilitators, and sometimes by participants themselves.

However, the relative scarcity of IST overall within the whole-group discussions and, when present, its fragility in the face of threats raise important questions about the viability of instructional rounds as professional learning practice. In particular, IST is threatened by synthesizing comments or comments about the process itself. Whether it is the articulation of a theme that synthesizes different interpretations of evidence or an evaluation of the instructional rounds process, these represents shifts away from the particulars of practice and toward the general.

What can researchers and practitioners learn from these findings? The implications for learning communities using instructional rounds and similar discourse-based professional learning practices are clear: Facilitators should explicitly and repeatedly press participants to provide specific evidence in their presentations of findings, as well as their comments and questions related to others’ findings. Probing for evidence (and disconfirming evidence) should become a staple of the facilitation of discussions during an instructional rounds visit.

The specificity that sustains IST is also a foundation of effective supervision of teachers, one of the three practices involved in managing instruction that Hallinger (2005) identified. Thus it is not just specificity for the sake of conducting rounds that is at stake, but the capacity of school leaders to make and defend claims about instruction using specific examples from a teacher’s practice if they are to enact instructional leadership in their schools and/or districts. Similarly, the specificity demanded by instructional rounds in focusing on student learning relates to two other practices Hallinger identified for managing instruction — coordinating the curriculum and monitoring student progress.

This continual press for evidence is essential but not sufficient. Practitioners might also consider how the use of IST could become more engrained in the fabric of instructional rounds visits, both culturally and practically. Culturally, a learning community might adopt norms that make explicit discourse moves such as supporting or challenging claims with specific examples from classroom observations. Participants might practice doing so through use of role-play activities and analysis of video recordings of whole-group discussions, much as the network in the study devoted early sessions to practice making descriptive rather than evaluative observations of classroom interactions.

The findings also highlight questions for further inquiry. Our study was limited by the unavailability of data from the small-group discussions that preceded the whole-group discussions. In these discussions, participants who observed classrooms as a team shared evidence and identified initial patterns and wonderings to share with the whole group. Because these discussions were not audio recorded, it is not possible to evaluate whether the smaller size and shared units of analysis allowed for more frequent or richer episodes of IST.

Researchers might also address other factors found to initiate IST, including questions related to the instructional core and disagreements of interpretations of evidence from classroom visits. While theses initiators were less frequent than direct invitations for specific evidence (usually by facilitators), understanding how they encourage forms of discourse generative of learning, such as instruction-specific talk, holds promise for adding power to the professional learning of administrators about key issues of instruction and student learning.

For school and district leaders to be instructional leaders, or leaders for learning, learning practices that develop understanding of the instructional core must become a staple of their professional lives. This is especially true given that many school leaders may enter administrative roles inadequately equipped in their understanding of the instructional core beyond that required of an individual classroom teacher (Brazer and Bauer, 2013). Instructional rounds offers a promising practice to support this ongoing professional learning; however, we must continue to unpack how rounds operates as a learning practice and how it can be enhanced to do so more effectively.