Contents

• Introduction
• Executive Summary
• Chapter 1: The Central Office in District-wide Teaching and Learning Improvement: The Importance of Transformation
• Chapter 2: Dimension 1: Learning-focused Partnerships with School Principals to Deepen Principals’ Instructional Leadership Practice
• Supporting Individual Principals in Strengthening Instructional Leadership
• Differentiating supports for principals’ instructional leadership consistently over the entire academic year
• Modeling Ways of Thinking and Acting
• Developing and Using Tools in One-on-one Assistance Relationships
• Brokering for Individual Principals
• Principal Networks: A Narrative Example
• An Example of Promising Network-based Assistance Practice
• Vignette 1: A Principals’ Network Meeting
• Unpacking the Example
• Summary
• Chapter 3: Dimension 2: Assistance to the Central Office–Principal Partnerships
• Chapter 4: Dimension 3 :Reorganizing and Reculturing All Other Central Office Units to Support Teaching and Learning Improvement
• Chapter 5: Dimension 4: Stewardship of the Overall Central Office Transformation Process
• Chapter 6: Dimension 5: Use of Evidence Throughout the Central Office to Support Continual Improvement of Work Practices and Relationships with Schools
• Chapter 7: What the Study Says about Central Offices and the System-wide Improvement of Teaching and Learning
• Methodological Appendix
• References

Click here to download the full report:
 Central Office Transformation for District-Wide Teaching and Learning Improvement

All three central offices organized their central office transformation strategies around new relationships between school principals and central office administrators to intentionally support teaching and learning improvement in schools, highlighted schematically in Figure 2 below.

Through these learning-focused partnerships, the central office aimed to make substantial investments in building the capacity of school principals to exercise instructional leadership. While definitions of instructional leadership varied somewhat across sites, the efforts across all sites suggested a consensus that when principals exercise instructional leadership, they work intensively with their teachers in and out of the classroom to critically examine the quality of their teaching practice and student work in an attempt to strengthen both. As noted earlier, various other central office administrators interacted with school principals in other ways and also directly with teachers. Nor were school principals the only people exercising instructional leadership within schools.⁶ However, a clear hallmark of the central office transformation efforts involved these dedicated central office administrators focusing on strengthening the capacity of school principals for instructional leadership.

We called these dedicated central office administrators, as a group, Instructional Leadership Directors (ILDs). As indicated in Table 3, the ILDs varied according to formal titles, backgrounds, resources, and supervisory relationships with school principals. For example, ILDs in Atlanta and Oakland all had prior leadership experience in education, including school principalships. By contrast, NYC network leaders came from inside and outside school systems and not all had been school principals.

All the ILDs reported that they were responsible for working with principals individually and in networks to strengthen principals’ instructional leadership. As one NYC network leader put it, “I help principals realize [that the] more they’re in classrooms, the easier their job gets.” Another explained that the main work involved engaging principals to talk “about how a lot of teachers work in isolation and don’t get the feedback necessary to grow. How teachers are losing professionalism because there is no true check [on their practice].” This network leader went on to explain that he/she helps principals understand that their main work as instructional leaders includes breaking down such teacher isolation through strategies that focus on feedback and helping teachers to grow. Many described such work as fundamentally involving helping some principals shift their conceptions of the principalship from a job involving mainly managerial functions to a profession centrally focused on helping teachers improve their teaching practice. As one ILD put it,

I … spend time in [schools] helping the principals … focus their work… . Working on the quality of teaching and learning. Looking at the student work. Looking at the rigor. Looking at best practices. Giving them feedback. [If I don’t] … it’s not going to pay out in dividends in the student achievement. Because … we are creatures of habit first of all. So taking a principal who has not spent time in their classrooms and getting them to shift their focus takes a lot of … intentional work. And then to be able to maintain that focus in a culture where people [e.g., teachers] are used to … keeping you in an office to deal with this one [student] all day—that’s a whole other level of work … . And then helping people [principals] to prioritize their time so that they do spend their time on the core business in the areas that matter the most.

However, we found clear differences in how the ILDs went about their work. Some ILDs frequently worked with school principals in ways that reflected the promising practices noted in Chapter 1. Because research across disciplines has identified those practices as important supports for helping others improve the quality of their own work, ILDs’ engagement in those practices can be thought of as a form of support that is likely to improve principals’ instructional leadership.⁷

Certain ILDs engaged in promising practices significantly more often than others over the course of the study period. Strikingly, ILDs who did so often or very often were also those whom principals almost unanimously reported supported their development as instructional leaders. Conversely, ILDs who infrequently engaged in these practices were mentioned by principals as having limited or negative effects on their instructional leadership. Reports from other central office administrators about more or less effective ILDs also confirmed these distinctions. On the basis of that evidence, we identify the following ILD practices in the one-onone principal relationships and principal networks as likely to help principals strengthen their capacity for engaging in instructional leadership:

• Differentiating supports for principals’ instructional leadership consistently over the entire academic year.
• Modeling ways of thinking and acting that reflected desirable instructional leadership practice.
• Developing and using tools.
• Brokering resources supportive of principals’ instructional leadership.
• Tapping all principals in a network as resources for each other around their instructional leadership practice, including providing opportunities for all, not only those in high-achieving schools, to take on leadership roles within the network.

We summarize our findings about each of these points below.

Supporting Individual Principals in Strengthening Instructional Leadership

Differentiating supports for principals’ instructional leadership consistently over the entire academic year

All the ILDs unanimously explained and demonstrated that they differentiated how they supported individual school principals in strengthening their instructional leadership. That is, they did not take a one-size-fits-all approach but rather worked with each principal differently depending on a variety of considerations. However, not all ILDs were able to sustain their differentiated work with individual principals over the entire academic year. Those who appeared inconsistent in this regard tended to be viewed by principals and other central office administrators as relatively ineffective in strengthening principals’ instructional leadership.

Two of our three interviews asked ILDs to identify the two-three principals they worked with most often and least often and to describe concrete examples of how they worked with those principals. Each interviewee provided examples of how they tailored their work with individual principals within and between these groups, based on a variety of factors including principals’ instructional leadership ability, school performance, and the extent to which the school principal reached out to them. For example, one NYC network leader described the range of the problems of practice in this way,

… it may be about sitting with their professional development team, listening to what they’re trying to put together, and then asking questions to help them through that. It could be in terms of looking at classes—an initiative that the school may have and they want to see how the instruction is going, or it could be because they want a different lens on a teacher that they feel is not performing up to par and they just want my input on that. It could be a parent meeting that they’re having to explain the data and how to look at the data, or things like strategies like how to read with your children, or building vocabulary—activities that they can do at home. It could be around having conversations with some principals that may be stressed and overwhelmed and talking crazy, like “I’m quitting.”

Likewise, in Atlanta, one School Reform Team (SRT) executive director spoke about the particular needs of a subset of new principals who had strong capacity to observe instruction, provide feedback to teachers, and plan for individual and group professional development; however, these particular principals lacked experience with operational issues which took time away from focusing on instruction. Accordingly, the SRT executive directors and principals agreed to focus their work together to improve the operational supports in those schools. By contrast, another principal had recently moved from a middle school to a high school. This principal had been a strong instructional leader the middle school but needed support to exercise leadership appropriate to the new high school setting. The same executive director focused work with that principal on using data to understand the dynamics of the school’s small learning communities and how to strengthen teaching and learning within those communities. Principals in these examples corroborated the ILDs’ accounts of how they addressed issues specific to the principals and their school.

Shadowing ILDs yielded further evidence of such differentiation. For example, on one occasion we observed an SRT Executive Director working with a new principal identified as in need of support around classroom observations. In the executive director’s words, “I recognize there is a learning curve [for new principals], but I need [the principal] to be focused on the right stuff [now, in observing in classrooms]. So, I can’t leave that to chance.” During the observation, the executive director stayed physically close to the principal through three classrooms and frequently engaged the principal in conversation about evidence from the classrooms that did or did not fit the standards of high-quality teaching outlined in an observation rubric. In a separate debriefing conversation, the executive director walked the principal point-by-point through the rubric and challenged the quality of the principal’s evidence and the fit with the rubric. The meeting concluded with the executive director listing specific actions for the principal to take when observing classrooms before the next meeting. By contrast, on a visit to the school of another principal, identified as more expert in the area of classroom observations, the executive director and principal observed the same classrooms with the rubric but with little to no dialog among them. During the debrief conversation, the executive director and principal compared teacher ratings, finding only a few discrepancies. The principal led most of the debrief, asking the executive director for advice on particular teachers and other issues related to school improvement plans. The meeting concluded with the principal sharing next steps for teacher development.

These and other examples suggest that when ILDs differentiated supports, they did not simply work with individual principals differently. Rather, they worked with each principal in ways that fit individual needs and strengths related to improving instructional leadership practice, much like expert classroom teachers differentiate instruction for individual students. One common strategy involved helping principals identify their own improvement goals or otherwise participate centrally in determining the focus of their work with their ILD. As one network leader reflected about the practice of network leaders in general,

What I think [all the Network Leaders do] really well is … we’ll work with you [i.e., school principals] around your questions [about teaching and learning improvement]. If you tell us what your questions are and you tell us what you want to work on, we’ll help you think that through, and we’ll help you figure out a way of pursuing that.

In Atlanta, ILDs’ engagement with principals in mutually identifying anchoring problems of practice evolved after many years of central office transformation. As one SRT executive director described, in the early years of the reform effort central office leaders chose an area of focus for the principals:

We started out with reading, and English, and language arts, knowing that that was our main focus because all of the data indicated that the students, if they couldn’t read, they couldn’t do anything else. And believe me, reading at the time was quite dismal. I think that the district average was 47% nine years ago in reading, and even my SRT was 37%. … The district is somewhere around 82% now, but I mean there’s some drastic gains out there. But the idea was that you had to fix the roof and also the sides, to go out the elementary schools first, because [that’s] the base.

This ILD went on to explain that as school performance improved, principals had the opportunity to take more ownership of instructional decisions, including having input into the focal problems of practice.

Now most of those schools have more autonomy through me. They’re still directed [by the central office] … but if they [principals] come up with something they want to do now, and they have the scores to justify it, [they can try something new].

These and other data suggest that when the ILDs worked with individual principals, they differentiated how they worked with them. However, some did not continue to work with individual principals in a differentiated way over the entire academic year. For example, some principals reported that they did not see their ILD very often. Others reported that their work with their ILD waned over the course of the academic year. For instance, one principal reported that meetings with the ILD decreased from monthly in the first semester to just one meeting in the second semester, and that despite reaching out to the ILD, “I rarely got responses.” When asked about the reason for the decrease, this principal reported, “I think it’s just like most things, particularly in education. … You get too many fires to put out, too many other priorities and so I just got the sense that [my ILD] had other priorities to deal with other than working directly with me or [my school].” But some veteran principals viewed the limited attention they received from their ILD as a sign that their ILD was differentiating support among new versus veteran principals. For instance, one principal reported,

[There are] … principals who have less experience than I do. They’re in their first year or their second year and I’m even now, with four years, in a different place than where they are. So, I think that they [the ILDs and other central office administrators] understand that and I think it directly impacts the way that they work with us. I think that sometimes it’s created some confusion in some ways because we’re not all necessarily asking for the same thing all the time, so I actually think on their end, it’s a rather difficult job to manage. It’s pretty analogous to having a class full of heterogeneous students where people need very, very different things. So, I think they have largely done a pretty damn good job of juggling the whole thing and trying to meet everyone’s needs. But, I think the demands are pretty great on them.

Some ILDs attributed their occasionally spotty focus on individualized support to resistance on the part of principals to keeping meetings with them and otherwise following up on their individualized improvement plans. As one ILD described,

… for too long, most of the principals I have … have been in their spots going on five years. And in that five years [in this district] there was no discussion about instruction—there wasn’t any, and they’ll all tell you that. None of them were evaluated so they didn’t get feedback, and so this is pretty new to them—have someone asking questions … . And they certainly aren’t used to someone contacting them at least once a week and interacting with someone at least once a week—not around instruction. There might have been someone calling and saying, “Where is your latest check-off sheet or something,” but definitely not instruction.

Multiple NYC respondents reported that some school principals had chosen the Empowerment Schools Organization as their School Support Organization precisely because they did not want central office involvement in their school, which may have contributed to this particular challenge in the NYC context.

In Oakland, the ILDs also struggled to maintain differentiated work with individual principals on instructional leadership due to excessive demands to work with principals on school compliance issues. Some NExOs resisted such demands. As one NExO described,

Last year I got completely awash in that logistical kind of side-tracking stuff. And so we as NExO’s made a commitment to 24 hours in schools focused on instruction every week. And so what I’m doing is I’m starting to ignore the noninstructional stuff … . Now, so if a principal calls me [with a non-instructional issue] I’ll act on it—[but if] it’s a cc on an e-mail… I don’t get involved … . And I don’t feel bad about it because I’m really getting feedback, too, from the principals [that we] … are truly making a difference for their instructional focus and what they’re doing for instruction for the kids.

However, others reported that by spring they had canceled most of their individual school visits due to personnel hearings and other matters.

Modeling Ways of Thinking and Acting

Modeling or demonstrating particular ways of thinking and acting are essential strategies for helping people such as school principals change their work practices. Models provide “an advanced organizer for the initial attempts to execute a complex skill … an interpretive structure for making sense of the feedback, hints, and connections from the master during interactive coaching sessions, … . And … an internalized guide for the period when the apprentice is engaged in relatively independent practice” (Collins et al., 2003, p.2). As one ILD put it, modeling or demonstrating instructional leadership actions proved a far more effective strategy for actually influencing principals’ practice than “telling them.” This administrator went on to say, “I recognize that there’s a delicate balance between what I know and what they need to know. And so telling them is really not an effective method.” Instead this person routinely modeled particular instructional leadership practices, “because, ultimately, when I leave, I want them to know how to do it [exercise instructional leadership].”

Research on learning also underscores that the strength of models as learning supports hinges, in part, on the modelers’ or others’ use of strategies that help the learner reflect on their learning (also referred to as “metacognitive” strategies), such as explicitly pointing out to principals what practices and ways of thinking the models are modeling and their underlying rationales for doing so (Brown & Campione, 1994; Collins, Brown, & Holum, 2003; Lee & Smith, 1995; Tharp & Gallimore, 1991). Absent such strategies, learners may not notice that a particular practice is being modeled or understand the underlying rationale for the modeling (Collins et al., 2003; Lee & Smith, 1995).

The ILDs varied in terms of whether and how frequently they used these forms of modeling as strategies for influencing principals’ instructional leadership practice. As noted earlier, ILDs who frequently modeled instructional leadership practices were most frequently identified by principals and other central office administrators as positively influencing principals’ instructional leadership. Conversely, ILDs who infrequently modeled tended to be seen by principals and other central office administrators as providing limited or ineffective supports for principals’ instructional leadership.

For example, ILDs frequently demonstrated for principals how to have challenging conversations with teachers about how to improve their teaching practice. For instance, several Oakland ILDs routinely modeled for their principals how to run meetings to help teaching staff make sense of data on how their instructional practice might be contributing to particular test results. One NExO facilitated staff meetings at three schools with each principal observing and documenting the strategies used in talking with teachers. The NExO and the principal later met to discuss the strategies, and the principal subsequently practiced those strategies in another meeting with teachers, with the NExO observing and providing feedback.

Similarly, at a meeting of New York City network leaders that we observed in the fall of 2007, one network leader described a school that almost received a failing grade on its school progress report. In a professional development session designed to address the school’s poor grade, the network leader reflected that everyone was “friendly”—meaning the teachers did not honestly confront the root causes of their school’s low-levels of performance. “We talked about that, that it’s okay to get friendly, but you got to get down to business.” What ultimately helped the school principal understand how to “get down to business” and have those challenging conversations was for the network leader to first model a direct and honest conversation focused on implications from the data for the principal’s personal leadership practice, and then to reflect with the principal on that conversation for ideas about how to have similar conversations with teachers.

ILDs also helped principals engage in instructional leadership when they modeled ways of thinking that reflected good instructional leadership practice. One NExO in Oakland described this modeling work as coaching the principal’s ways of thinking:

… talking through what [the principal’s] thinking is and then helping him to see where that might take him, so the principal has time to stop and actually think about why he is making the decisions that he’s making … . [A principal will oftentimes jump to decision-making without stepping back and really thinking about how he’s making [decisions], who he’s involving in the process, and then what are the consequences of that … .⁸

The NExO went on to illustrate this process of modeling thinking with a specific example that we corroborated in interviews with the principal,

So [a principal] asks staff to do something … . she may say, “I want you all to make sure that you post objectives … [in your classrooms], and expects once [this has been] said, that it is going to happen…. And one of the teachers says “No, I don’t want to do that,” or a teacher doesn’t do it and then … [the principal’s] immediate reaction is “Okay, we’ve got to do something to get this teacher out of here.”… Without understanding … the culture of [the] school. [The principal] doesn’t necessarily put herself in the position of those she’s attempting to move … . [So my work involved] taking her back to where the initial thought happens so that she can understand there are other options, and why is she thinking that, what about her and her approach to the work is taking her there?

Through the metacognitive conversations, the ILD provided the principal with specific examples of other ways to think about the situation, consistent with instructional leadership, and how to weigh the pros and cons of each choice.

The ILDs whose work was corroborated by positive reports of supporting principals’ instructional leadership not only modeled ways of thinking but also routinely used reflective strategies to help principals see what they were modeling and why. As one Oakland NExO explained, “If I’m going to have any impact at all on these schools, I have to … teach them [my principals] and teach them why we’re doing what we’re doing…to help them to become instructional leaders.” This ILD further explained that unless the principals understood the underlying rationale for certain practices, they were more likely to perceive their engagement with the ILD as a directive and evaluative rather than supportive relationship and to resist participating in it. We confirmed most of these reports through direct observations. For example, during one observation an Atlanta ILD demonstrated for a school principal how to use data to provide feedback to a particular teacher. The ILD then reflected back to the school principal what kinds of strategies he was using to help the teacher see the urgency to change her practice, including displaying data in certain ways and asking the teacher first to interpret the data.

Counter examples—instances in which ILDs did not model ways of thinking and acting for principals but rather stepped in and essentially did the principals’ work for them or told them what to do without creating intentional opportunities for their principals to observe or practice the work—confirmed the potential power of reflective modeling. While such activities resulted in some principals’ work getting done, they did not seem in any way tied to building principals’ capacity for instructional leadership (or other activities, for that matter). For example, when asked how she helped a school principal fund a particular instructional program, an ILD described opening up a school budget and making adjustments herself. When asked directly to what extent she also demonstrated these budgeting strategies to her principals or thought such demonstrations might be important for the principals’ development, this ILD responded that it was easier for her to just go in and make the change. Similarly, an ILD believed that certain principals were not using school-based coaches in effective ways. Rather than modeling alternatives or otherwise engaging principals themselves in addressing this challenge, the ILD stepped in for principals and gave direct mandates to the coaches about how they used their time. Perhaps not surprisingly, the principals in these examples reported that their ILDs did not help them appreciably build their instructional leadership capacity.

Developing and Using Tools in One-on-one Assistance Relationships

All the ILDs across all three districts used various materials in their work with individual principals.⁹ Some of these materials appeared to be what learning research distinguishes as “tools”—materials intentionally designed with features to engage principals in new ways of thinking and acting that reflect good instructional leadership practice. Conceptual tools include “principles, frameworks, and ideas” (Grossman, Smagorinsky, & Valencia, 1999). These tools generally function to frame how people think about issues as a main strategy for shifting how people think and act. Practical tools likewise aim to shift people’s thinking and actions but by leading with specific examples of “practices, strategies, and resources” that have “local and immediate utility” (Grossman et al., 1999, p. 13–14). While conceptual tools are meant to shape decisions across multiple activity settings, practical tools are generally constructed around a particular activity setting. Across both tool types, the effectiveness of a tool in influencing users’ practice hinges substantially on the extent to which people use tools to engage learners in challenging conversations about the implications for their own practice of the ideas embedded in the tools (Honig & Ikemoto, 2008).

We found four types of tools commonly used by ILDs in their assistance relationships with individual principals in at least two of our three study districts:¹⁰

• teaching and learning frameworks intended to define common conceptual understanding of what constitutes quality teaching and learning (Atlanta and New York City);
• school walkthroughs and other classroom-observation protocols that guided how leaders observed classroom teaching and learning (Atlanta, New York City and Oakland);
• cycle-of-inquiry protocols (New York City and Oakland); and
• data-based protocols designed to focus principals’ instructional leadership practice on outcomes of various kinds (Atlanta, Oakland and New York City).

All the ILDs in the three systems incorporated tools into their work with school principals but not all of them used them to engage principals in challenging conversations in the ways that research suggests holds promise for strengthening principals’ instructional leadership. Consistent with the other findings, ILDs who did engage principals in such conversations were also those ILDs who received positive reviews from other central office administrators and school principals themselves about the quality of their work with school principals.

Teaching and learning frameworks. We observed ILDs in two of the three districts using instructional frameworks—conceptual tools designed to help define “quality” teaching and learning. This practice was most common in Atlanta, where all the School Reform Team (SRT) executive directors used a districtsponsored, externally developed tool—the 26 Best Practices—in their work with individual school principals. The tool, derived from research on and experience with teaching for conceptual understanding, called on teachers, among other practices, to routinely ask students questions that moved beyond simple recall and required students “to think, synthesize, evaluate, and conclude.”¹¹ We documented how the SRT executive directors used this tool often in their discussions with individual principals to critically examine the quality of classroom teaching. An SRT executive director explained in a meeting with colleagues, “At one time we were all using different instruments” that defined and measured high-quality teaching and learning. Now with the 26 Best Practices,

We have a centralized … instrument that all of us use—whether you’re a principal, teacher, or central office executive director—that we created to improve the principals. And that’s the [basis for the] audit that we use when we talk about their performance … . So everyone is familiar with that.

Observations and interviews confirmed that principals’ work with SRT executive directors using this tool enhanced their ability to provide targeted feedback to teachers about their instruction. In a typical comment, one Atlanta principal said, “I believe that I have been able to provide my teachers with great feedback from that instrument because it is measuring their performance to a standard, to a real rubric they can see.” They also showed us written records of feedback to individual teachers with evidence from classroom observations organized around the 26 Best Practices.

By contrast, the absence of such a tool in Oakland seemed to compromise the ILDs’ work in strengthening principals’ instructional leadership. For example, we observed that when some Oakland NExOs accompanied principals into classrooms to look critically at the quality of teaching, their discussions were not rooted in a similarly clear, consistent conception of high-quality teaching and learning. Perhaps as a result, the discussions about the quality of teaching on these visits often remained at a superficial level, with principals noticing whether or not objectives were visibly posted in the classroom for example but not teachers’ actual pedagogical strategies or interactions with students, and NExOs generally not responding with challenging questions about principals’ understanding of highquality teaching and implications for principals’ practice. Likewise NExOs for the most part reported that some of their principals lacked knowledge of high-quality teaching and learning. Some other central office administrators similarly noted that ILDs were not operating from a common research-based framework defining such classroom practice.

School walk-throughs.¹² In all three of our sites, ILDs routinely relied on protocols that called for principals to observe teachers’ classroom practice to collect evidence of that practice for use in subsequent conversations with the teachers about how they might improve their teaching practice. NExOs in Oakland, for example, worked together with their Chief Academic Officer to develop a master walkthrough protocol that they each adapted for use with different school principals. Four of the five NExOs in our study used the protocol as the basis for monthly meetings with each school principal. In Atlanta, the SRT executive directors translated the 26 Best Practices instrument into a protocol for classroom observations.

The adapted tools included a lesson observation form, which provided space for observers to record their rankings of the intensity-level of various practices and any instances of students being off-task, and a form to use to guide feedback to teachers. In NYC a subset of network leaders regularly observed classrooms with individual school principals and in most cases used a formal protocol as a tool for anchoring those observations.

However, ILDs’ use of these protocols to anchor challenging conversations with principals about their instructional leadership practice varied greatly. As one principal recounted, the ILD conducted walkthroughs but the walkthroughs did not have a positive influence on the principal’s instructional leadership practice. The ILD would …

… just come in [to the classroom], stand at the back, take some notes, walk away, and send me an e-mail a couple of days later and say how horrible the observation was. If it’s really that bad then you should have been compelled to have a conversation with that teacher or at least a conversation with me. “Hey … this is what I just saw—let’s go into the class together this next period, observe it together and find out where we can help support this teacher and improve his instruction.”

Cycle-of-inquiry protocols. In New York City and Oakland, ILDs used formal cycle-of-inquiry protocols to ground one-on-one interactions with principals around instructional leadership practice. Some respondents in our study used the term “inquiry” to refer to any strategy that involved posing questions of principals. By contrast, cycle-of-inquiry protocols have been associated with helping improve principals’ (and other professionals’) work practices when they prompt principals to identify a specific problem of practice related to their efforts to improve teaching and learning; to collect evidence to help them better understand the underlying causes of that problem; to develop strategies supported by a rationale for how the course of action would address the problem; and finally, to continually collect evidence to assess progress toward solving the problem of practice (Copland, 2003). One of the ILDs in Oakland described the importance of the cycle of inquiry process in this way:

… if either of them [two principals] get good at that cycle of inquiry or how they look at new data in enhancing staff skills about instruction [as the main focus of one of their inquiry cycles], and we go really deep into that, then those systems and that discipline will have larger effects into the rest of their work.

In New York City, a district-wide initiative called the Children First Intensive (CFI) required all schools to convene school-based teams in a year-long cycle-of-inquiry process around the progress of a subset of students at each school.¹³ Some New York network leaders said they adapted the CFI tools and processes to help engage principals in critical conversations about how their practice mattered to improving results for those students. In the words of one,

So now we’re working with assistant principals and principals and teachers to really own this [Children First Intensive and other assessments] and make it relevant for them…. [T]hat’s the work. [T]hat’s the focus for us. Not walking around a building and making recommendations, but more okay, how are you using the tools of accountability, how are we using the inquiry team, how are we helping them identify a school-wide focus, identify a population of kids? So that’s the instructional work that we do.

Likewise, we observed how at least in Oakland, those NExOs identified as supporting principals’ instructional leadership used cycle-of-inquiry protocols to engage principals in challenging conversations about their instructional leadership practice. For example, we saw e-mail exchanges between those NExOs and their school principals detailing action steps the NExOs and the principals had agreed the principals would take as part of a particular cycle of inquiry process, and products the principals would review at upcoming meetings with their NExO as evidence of their work.

Data-based protocols. All the tools above engaged principals in considering various data related to the quality of classroom teaching and learning as fundamental to their exercise of instructional leadership. Some ILDs across our districts did not simply put raw data in front of school principals but rather engaged school principals in data-based protocols—tools that presented specific data and posed questions about the implications of the data for principals’ instructional leadership. Principals’ whose ILDs utilized such protocols were more likely than other principals to report and demonstrate that they understood how to use student performance data and data from classroom observations to inform their own leadership practice.

For example, in Atlanta, we frequently observed SRT executive directors using the 26 Best Practices to organize evidence they collected while reviewing student work and teachers practice along with different “dashboards” for displaying student performance data. Principals generally reported that they knew what their SRT executive directors were looking for when they visited their schools and the kind of questions they would ask about action on the results. In a comment typical across almost all of the ILDs in this district, one reported that such tools helped them juxtapose seemingly contradictory evidence of principals’ work in ways that challenged their leadership actions in support of improving classroom teaching,

I had one principal—every one of the teachers got 100% on their performance evaluations … . They [the teachers] only have about 57% of their kids meeting or exceeding the state standards. A third of their kids didn’t pass the … [state test]. When … every one of the teachers got 100% on their performance evaluation [from the principal], I said “Who’s 100%? You? Who? How does everybody get 100%?” I started asking principals to do a correlation between the performance evaluation ratings you’ve given these people and the data. How can you have 90% of your staff exceed and 40% of your kids are exceeding. I don’t understand the correlation there … . You can’t just come in here and do a mediocre job for our kids anymore. You got to get better.

This executive director went on to describe one principal’s reaction to working with data in this way as typical of many principals’ responses,

[The principal] told me … “[y]ou forced my hand. You have forced me to really understand this and take a look at it and really see. I get it.”… And so [this principal] used the words that I “forced him.” But it will not be a force in the future because [this principal now] gets it … . [W]e were having the conversations …, but I had to find some way in black and white and actually make [the principal] work through it.

In significant counter examples, some ILDs engaged their principals with data but not as part of a protocol for helping principals meaningfully grapple with implications of the data for their own instructional leadership practice. Such examples typically featured the ILDs presenting principals with data to help justify why principals should follow the ILD’s directives when making particular decisions.

Brokering for Individual Principals

ILDs also helped advance individual principals’ instructional leadership practice when they engaged in brokering activities. Research on learning emphasizes the importance of brokering to helping improve the learners’ own and others’ work practices. Brokering involves both bridging people to new ideas, understandings, and other resources that can potentially advance their work, and also buffering them from potentially unproductive external interference (Wenger, 1998). Often called “boundary spanning,” such activities contribute to improvement when they involve people such as ILDs not simply passing along new resources but translating them into forms that others may be especially likely to use (Aldrich & Herker, 1977; Cobb & Bowers, 1999; Dollinger, 1984; Tushman, 1977; Tushman & Katz, 1980). We distinguished more and less promising bridging and buffering activities based on the extent to which they seemed to increase the time that school principals spent on instructional matters.

Bridging. ILDs in all three districts supported their individual principals by bridging them to resources beyond those immediately available in their one-onone relationship. First, ILDs connected school principals to others in their central offices who supported their instructional leadership. In two of our districts, these resources very frequently included members of the network teams who worked for the ILDs—model teacher leaders (MTLs) who were part of each school reform team in Atlanta and achievement coaches who were on the staff of each network team in New York City—who engaged with principals and schools daily around instructional matters. In Atlanta in particular, the MTLs frequently helped principals in various ways, including collecting and organizing school progress data, modeling how to facilitate school leadership team meetings around issues of teaching and learning improvement (rather than, for example, mainly governance or operations), and modeling how to provide meaningful feedback to classroom teachers. On some occasions, the executive directors assigned their entire team of MTLs (between 10 and 14 people, typically) to struggling schools to assist with various dimensions of school improvement including principals’ instructional leadership. As one executive director described in one instance,

My whole team has been assigned to support [a school] this year. We went in that building in September. I had a principal that’s been on the job for about a year. I started hearing rumblings in August that … kids were just kind of not focused, teachers not focused. So … the first week of September we did a school kind of blitz site visit and spent about two and a half hours in there … going into classrooms spending 20 minutes … looking for evidence of teaching and learning, gave the principal some feedback, went back about three weeks later. And I realized that they had gotten off to a rocky, rocky start … . I told everybody on the team that … I expect everybody to go into that school once a week and visit classrooms, observe instruction, give feedback, provide support to teachers.

During our observations in Atlanta we almost always found these staff working in classrooms in these ways.

The Oakland NExOs did not have staff to deploy to principals and schools but some of them frequently tried to connect principals with others in the central office who served as vital instructional support resources. Two NExOs, in particular, routinely invited other central office administrators to attend their individual school site visits and arranged for staff from the district’s instructional services unit to provide modeling for principals in targeted subject areas. However, NExOs’ attempts at bridging to other central office administrators for instructional resources did not always result in increased resources to principals. As one NExO described,

So what I do is I send my calendar around and [a central office administrator] says “Oh, I want to come to schools with you,” and I say “Just let me know anytime” and I send my calendar, but she tends not to contact me … I think [that this is because] she just feels like she has her] own work. She is busy. I mean we do make arrangements for these guys to get in when we’re doing these … school walk-throughs and I do really enjoy listening to them, having them be part of the debrief, but I think it would be great if they were there more often.

ILDs also linked to others in the central office to help principals with activities traditionally distinguished as operational, managerial, or otherwise non-instructional in nature. ILDs cast such bridging activities as essential to helping principals spend less time on those issues and more on instructional matters. Typically ILDs became involved in these non-instructional matters when other central office units did not respond to principals in a timely manner or with the right services. ILDs found such intervention important in all our districts, including Atlanta, where the central office had been working the longest to improve the quality and responsiveness of central office units to principals. As one executive director in that district reported,

Part of our work as SRTs is to broker services and support [for school principals]. I don’t feel like there’s any [central office] department that I don’t have contact with. Quite frankly, we have some departments I wish I didn’t have as much contact with … I mean … a manager and 35 [staff work in that unit] but when there’s a problem in a school [related to that unit] sometimes the principals feel like they have to get me involved so they get an immediate resolution And I really wish it were not that way because I need for departments and individuals to be as responsive when a principal calls the first time as they are when I call … . I probably talk to my [other central office] person every day, multiple times a day. Usually I’m calling to either … get her to answer a question about a [unit-specific] issue or [other unit-specific] issue on behalf of one of the principals so that they can focus on their school work.

ILDs also connected principals to resources outside the district system to help improve principals’ instructional leadership. Such efforts typically involved not only identifying the resources and establishing relationships with outside individuals and organizations but also raising or redirecting funds to help principals pay for those resources. For instance, an Atlanta executive director described efforts to bridge to external resources as looking for resources for principals “anywhere … . I don’t think you could have any boundaries to where you can get the services.”

By comparison, in Oakland, NExOs’ activities to help principals access outside resources tended to involve identifying funds schools could use to purchase certain services and then encouraging schools to use those funds in particular ways. Such fundraising and advocacy with principals around spending seemed especially important in this district given severe budget shortages within the district and statewide in California, NExOs’ own extremely limited budgets, and Oakland’s Results-Based Budgeting system that aimed to maximize the dollars under principals’ rather than central office discretion.

Buffering. In addition to these various bridging strategies, ILDs buffered their principals from external demands as a strategy to support their instructional leadership practice. We distinguished buffering activities as those that protected or shielded individual principals from demands that jeopardized their ability to focus on instructional leadership work. Specific forms of buffering across all three districts included taking demands off principals’ plates, standing in for other central office units so the principals interacted with ILDs rather than those other units, and translating external demands to limit the amount of time principals spent sorting through and making sense of them.

First, ILDs in all three districts occasionally took demands off principals’ plates either by deciding that principals did not have to participate in particular activities, such as district-sponsored informational meetings, that might require the principal to be away from school or otherwise not focused on instructional issues, or by meeting those demands themselves. The latter activities differed from ILDs efforts to substitute for principals, discussed above under modeling, in that they involved ILDs running interference with activities that did not seem essential for principals to engage in as part of taking responsibility for instructional improvement efforts.

Every single one of the executive directors in Atlanta reported in interviews that they had been specifically charged with taking demands off principals’ plates when doing so would help focus efforts on instructional leadership. A New York City network leader put it this way:

So, one of the things that we’re responsible for is taking away many of the distracters that stop schools from being able to focus on teaching and learning and that’s what we’re trying to do. We take away those distracters, then they don’t have those time consuming things, you know, that stop them from really focusing in on instruction.

ILDs also stood in for, and did the work of, other central office units to improve the quality and relevance of supports the principals received. For instance, executive directors in Atlanta reported that their job included stepping in and taking care of issues for school principals if other central office staff did not respond and the principal brought the issue to their attention. As one executive director said, “When a principal or an AP reports a problem to me or to my office, it is our job to take care of that. I … take care of it.” This person went on to describe an incident where a principal had to take time finding a vendor to remove sexually explicit graffiti from the school building because the central office staff responsible for such work had not responded to the principal’s request for assistance in a timely manner. “I was so angry that I had a principal that had to deal with that because that’s not what they’re supposed to be doing.” In Oakland, some NExOs likewise took on the work of other central office units to help principals avoid interacting with non-responsive central office staff. As one principal confirmed when describing how one central office administrator never returns their calls: “[My NExO is] who I call. Period.” Network leaders in New York City similarly reported that a key part of their work involved stepping in for other central office units to improve the quality and relevance of the services available for their individual schools’ teaching and learning improvement efforts.

In New York City and Oakland, the ILDs buffered school principals from multiple and sometimes counter-productive external demands, especially from the rest of the central office, by translating those demands into forms that took less of principals’ time away from their instructional leadership. As one NExO described these activities,

I’m a buffer and a translator. I need to take mandates, expectations and re-frame them in such a way that they are meaningful and relevant and manageable for principals. That’s my job. And to break it down for them and to simplify and tell them stuff that, especially for my new principals, everything is not equally important, but there are some things that [I say] “Don’t drop the ball on this. Don’t you dare miss a teacher evaluation deadline.” Those are just [too important].

In other words, this administrator and some other ILDs did not simply pass on external demands to school principals, but either helped principals understand how to engage in them in ways that reinforced their own efforts to focus on teaching and learning improvement or suggesting that principals limit the time they spent on those demands. One NExO described helping principals not to simply comply with district-mandated accountability requirements but to adapt how they implemented those mandates to reinforce their efforts:

The school site plan, SPSA [Single Plan for Student Achievement]… [had a ] brand new format, brand new expectations, brand new template last year... . So I set an expectation number one that when [my principals are] doing this [plan development] that they would be very clear and deliberate [about] what were the gap areas that they were filling based on what data. And that they would have an instructional practice focus in their school. They could have more than one, but they would have at least one. So I basically took a district mandate and made it [relevant to principals].

Translation also involved repackaging other information from the rest of the central office so that it would be easier for principals to understand and address. As a principal described, “[My NExO] has this sort of summary e-mail of all the things we need to do that week, but then she forwards on all the other e-mails that I’ve already gotten but just like puts her little spin on it, like ‘you really should read this’.” In another principal’s words, “[My NExO] sends emails pretty much every week: updates, forwarding emails from the district, sending emails regarding deliverables that are due as reminders.” A similar pattern prevailed in New York City: All the network leaders reported filtering how their schools receive information from the central office.

Principal Networks: A Narrative Example

Those practices—modeling, tool development and use, and brokering—also helped us identify principal networks that appeared to offer promising supports for principals’ instructional leadership.¹⁴ ILDs in these networks also created opportunities for all their school principals to serve as resources for each other, regardless of their starting capacity. That practice reflects the research-based notion of “legitimizing peripheral participation”—a set of practices in high-quality learning environments that help learners improve their practice by seeing themselves as valuable members of professional communities regardless of their level of knowledge or skill and as on a trajectory toward improving their performance and therefore capable of improvement (Lave & Wenger, 1991). In such communities, even novice members of a group come to strengthen their practice not by sitting on the sidelines and observing but by jumping in even if in modest ways.

These distinctions between networks were not subtle but rather distinctly clear in our data. For instance, one principal described meetings that did not reflect these features as “not effective” in getting the principals to “calibrate their thinking about high-quality teaching and learning.” Another principal reported, “I’m not a sit and get person and it [the network meeting] seems more like sit and get. So every time that I come here … it’s drudgery . … It should not be that I dread it [the principal network meetings].”

An Example of Promising Network-based Assistance Practice

In the extended narrative example that follows (Vignette 1), derived from our data but edited to protect confidentiality, we illustrate these promising network practices as they played out in one context. We then briefly discuss how the narrative example illustrates modeling, tool use, brokering, and opportunities for principals to serve as resources for the network.

Unpacking the Example

This composite case represents a number of aspects of ILD practice in networks that appear promising for strengthening principals’ instructional leadership.

Modeling in networks. In the case above, Dr. Jones modeled the importance of principals interacting with students during classroom observations as a key strategy for understanding the quality of teaching in a classroom. Using metacognitive strategies, she also called principals’ attention to the fact that she was modeling how to sit with students and engage with them during classroom observations and gave the principals time to discuss the underlying rationale for such strategies and their overall importance. Though not visible from the one meeting described above, Dr. Jones also modeled how to run a meeting focused on deepening participants’ knowledge of high-quality instruction, as a way of helping principals run such meetings for their own teachers. At meetings earlier in the year, we had observed Dr. Jones leading the majority of the network meetings to demonstrate how to facilitate discussions about high-quality teaching and learning. In most subsequent network meetings, such as the one described above, Dr. Jones intentionally supported principals taking over meetings themselves to practice such meeting facilitation. Consistent with high-quality modeling practices, Dr. Jones met with Ms. Rosario before the meeting to prepare and after the meeting to debrief the quality of Ms. Rosario’s performance. Jones also modeled thinking in her work with Ms. Rosario by talking with her before the meeting about the importance of thinking about the walkthrough template as a device for focusing principals on a few aspects of instruction.

In counter examples, ILDs seemed more often to direct principals’ practice rather than model it. For example, at one such network meeting, a small group of principals was grappling with the pros and cons of different approaches to improving instructional rigor at one school. The ILD of this network, while walking by the group, overheard part of the discussion and interrupted with specific instructions for what the principal should do. Subsequently, the principals stopped working together to make sense of which improvement approach to pursue and carried on several side conversations.

Developing and using tools. In the case above, Dr. Jones did not simply present a summary of best testing practices but developed a set of materials and a process for engaging principals in challenging conversations about those practices and how principals might use them in their own schools. In addition, Dr. Jones used the Learning Walk protocol template prior to the meeting to engage Ms. Rosario in challenging conversations about how to help principals learn from the schools’ strengths but also to provide her with critical feedback to strengthen her own performance. During the meeting, the Learning Walk protocol prompted principals to look for specific evidence of particular teaching practices. Reinforcing those dimensions of the tool during the debrief conversation, Dr. Jones challenged principals not simply to provide summary evaluative comments about the classroom but to point out specific evidence supporting their claims. We also observed counter examples, in which the ILDs used so many tools during network meetings that principals treated those tools as paperwork they had to complete rather than as materials to aid in their own learning. In other instances, while using a Learning Walk protocol in their meetings, the ILDs only asked principals to report “wows” and “wonders” without pushing them to provide evidence of their observations or rationales for identifying certain classroom teaching practices as positive or negative.

Brokering in networks. Dr. Jones bridged the principal network to the assessment office in an effort to capitalize on expertise in that office relevant to a common problem of practice in the network: improving teaching and learning for struggling ELL students. The assessment director did not simply show up and give a perfunctory, generic quarterly departmental report; rather, Jones worked with the director in advance to help tailor (or translate) the presentation for the network principals. Dr. Jones also connected the principals to non-instructional resources. However, she intentionally and strategically limited the amount of network time spent on such matters to brief informational items in the last 15 minutes of the meeting. The latter involved substantial work on Dr. Jones’ part to communicate with other central office administrators about why she was denying their request to address her principals directly and otherwise take up network meeting time with issues she did not see as central to their focus and which could be covered with a handout. In network meetings run by some other ILDs, time became consumed with presentations by other central office administrators about central office mandates. Notice, too, that in presenting the new policy about how to file for a curriculum exemption, Jones did not simply reiterate the central office policy. Rather, she “translated” the new policy by peppering her presentation with specific advice about how her principals might have particular success in securing a waiver.

Creating opportunities for all principals to participate in their network as resources. Dr. Jones involved Ms. Rosario, the host principal, in leading key substantive parts of the meeting, showcasing her particular strengths in working with English language learners. Dr. Jones similarly used network meetings at other schools and work with host principals in advance to feature strong or promising practice at each network school and how each principal might serve as a resource for other principals in the network. As the comments from the principals in the case above suggest, such a strategy successfully identified Ms. Rosario as a network resource: other principals asked if the ILD could create further opportunities for the visiting principals to learn from Ms. Rosario. At the same time, the meeting process Jones and Rosario co-designed did not stop at “show and tell”; rather, the principal focused her visiting colleagues’ classroom observations on aspects of her school she aimed to improve, thereby putting her own school and her instructional leadership practice on display. In this way, the visiting colleagues were invited not just to listen passively and perhaps share stories of how they shared her struggles, but to become co-investigators in the work of improvement, with the assumption that all of the principals in the network were capable of providing her with some useful advice.

Jones also intentionally grouped her principals into pairs during one portion of the meeting to match principals with particular strengths and needs. As Jones commented, such strategies were “just like organizing a classroom” where you differentiate groups of students in particular ways, depending on the activity, to provide members opportunities to teach and learn from others. Other ILDs, who were less adept at engaging all principals as a resource to each other. regularly identified the same principals as always high or low performing, sometimes suggesting that certain principals could just as well sit-out certain network activities, rather than distinguishing how certain principals had strengths in different domains and reinforcing the importance of all principals participating in network activities. Other ILDs rarely if ever called on principals to exercise leadership within their networks.

Summary

In sum, despite the different histories and conditions in our three focal central offices, all three organized their central office transformation efforts around new relationships with school principals focused on strengthening principals’ instructional leadership. All three systems created new central office positions specifically focused on such support. How these ILDs went about their work varied in ways that, based on both prior research and study data, would make it more or less likely for them to positively influence principals’ practice. Certain ILD practices seemed particularly promising for strengthening principals’ instructional leadership: consistent differentiation of support for individual principals, modeling ways of thinking and acting consistent with principals’ instructional leadership, and using tools and brokering resources. Creating opportunities for principals in networks to serve as learning resources for each other also appeared promising.

However, these ILDs did not go it alone. We found that their ability to engage in this work with individual school principals and principal networks depended on support from other central office administrators. We explore such support in the following chapter.

< < Previous | Next > >

References

6. As a companion study documents more fully (Portin et al. 2009), a growing cadre of teacher leaders and other staff in many of the schools we studied were engaged in instructional leadership activity, often in teams led by the principal.

7. One practice noted in Chapter 1—differentiating supports for principals’ instructional leadership consistently over the entire academic year—was not anticipated by the learning theories on which we based our study, but emerged as a clear pattern distinguishing how ILDs worked with principals.

8. To maintain the confidentiality of the participants in this study, we alternately use “him” and “her” in direct and indirect quotes.

9. School principals in our districts had various tools available to them from different sources, including central office administrators other than the ILDs and external organizations. Here, in keeping with our study focus, we specifically address tools the ILDs used in the context of their one-on-one assistance relationships with principals.

10. Honig & Ikemoto (2008) distinguish these tools as organizational tools, those commonly in use across organizations.

11. Atlanta Public Schools (N.D.). Lesson Observation Form. Atlanta, GA: Author. Collected in 2007.

12. These tools go by various names in practice including Learning Walks. Since the term LearningWalk is copyrighted by the Institute for Learning, we use the more general term, “walk through” to refer to these tools. However, some of the network leaders in NYC specifically used the Institute for Learning’s LearningWalk protocol which had been developed through consultation with former NYC Community School District #2 Superintendent Anthony Alvarado.

13. New York City Department of Education. (2008). Children First Intensive Inquiry Team Handbook. New York City: Author.

14. Study data about the network relationships ranged widely from: Interviews with network leaders and school principals in NYC; interviews with various central office staff and occasional observations in Atlanta; and multiple interviews and regular direct observations of the network meetings of five of the eight NExOs in Oakland over time. Because we had substantially more network meeting data from Oakland than the other districts, findings from this site were our strongest influence on how to organize this subsection.