Making the most of instructional coaches



Although coaching shows promise for professional development, some instructional coaches are spread too thin to focus on instruction.


Instructional coaching is among the fastest-growing forms of support for teachers’ professional learning (Darling-Hammond et al., 2009), and for good reason. When compared to other common forms of teacher professional development, such as one-day workshops, coaching is an exciting possibility because it embodies two essential aspects of effective professional development: It is ongoing and located in the context of teachers’ daily work (Hawley & Valli, 1999). Thus, teachers can try out and adjust new instructional approaches in their own classrooms, with their own students, and with the support of someone more knowledgeable than themselves (Cobb et al., 2018; Kane, 2016). This ongoing, site-specific support is a near-holy grail for teachers’ professional learning.

But despite coaching’s promise, the evidence about its overall effectiveness is inconsistent. Although many studies report that coaching initiatives support students’ academic improvement or teachers’ development of new instructional practices (Sailors & Shanklin, 2010), others report mixed results (e.g., Bean et al., 2010). Why is this? One reason may be that coaches’ job descriptions often include a wide variety of disparate duties, which erodes the time coaches have to work directly with teachers on instruction. Studies show that coaches may spend only about one-quarter to one-third of their time working with teachers to improve instruction (Bean et al., 2010). They spend the rest of their time on a multitude of other duties, such as locating curricula, tutoring students, substitute teaching, collating test data, or organizing students’ log-in information for various software programs (e.g., Kane, Cobb, & Gibbons, 2018; Kane & Rosenquist, 2018).

Coaches’ lack of time working with teachers to improve instruction is unfortunate, since coaches can be an important source of content-specific instructional expertise in schools. Plus, coaches’ time is expensive: By one estimate, coaching costs 12 times more than traditional, workshop-style professional development (Knight, 2012). The cost is worth it if coaching can help teachers learn, use, and sustain new instructional practices in ways that traditional “one shot” workshops cannot (Hawley & Valli, 1999). But to achieve their potential, coaching initiatives must be designed to maximize the time coaches spend working with teachers to improve instruction. In this time of strapped budgets, careful planning and thoughtful policies around coaching initiatives are particularly important since ineffective coaching initiatives may be short-lived.

Investigating coaching initiatives

This analysis on which this article is based was undertaken as part of a larger project called Mathematics and the Institutional Settings of Teaching (MIST) that investigated how district-level systems support teachers’ professional learning. Over eight years, we collected hundreds of hours of interviews with district leaders, school leaders, instructional coaches, and mathematics teachers, as well as survey data, videotapes of teachers’ classroom teaching, and assessments of teachers and coaches’ understanding of mathematics and mathematics pedagogy.

One area of interest was how district- and school-level policies and expectations were related to the time instructional coaches spent working directly with teachers on issues of instruction (Kane et al., 2018; Kane & Rosenquist, 2018). We used quantitative survey data to ask 23 middle school mathematics coaches across our original four partner districts to report the percentage of time they spent participating in a variety of duties commonly assigned to coaches (e.g., co-teaching, analyzing data, making copies). Our goal was to understand the amount of time each coach spent in activities that research suggests are potentially useful for instructional improvement, such as modeling or analyzing classroom video, versus the amount of time they spent in duties unlikely to support instructional improvement, such as making copies or substitute teaching (Gibbons & Cobb, 2017).

Our next step was to conduct a close qualitative analysis of 72 interviews conducted over two years with district leaders, school leaders, and mathematics coaches in one of the partner districts. (Unfortunately, data from teachers were not available.) This district’s policies around coaching changed substantially during those years, as did coaches’ reports of their time use, which gave us more information to work with.

The coaches in the study were either hired by a district mathematics specialist or by an individual school’s principal. District-hired coaches were typically expected to spend four days a week at three to five schools. They spent Fridays at the district office, receiving professional development or doing other district work. This meant that district-hired coaches were in any given school building once a week, at most. By contrast, school-hired coaches worked at only one school, five days a week. As seen in other research on coaching, coaches in our study faced broad expectations for their work, and some, but not all, expressed frustration about the lack of time they had to work with teachers on issues of instruction. Coaches’ use of time was most closely related to whether they were hired by the district or the school.

District-hired or school-hired

Before beginning this analysis, we assumed that school-hired coaches would be better able to work with teachers in a consistent, ongoing way, because they would be in the building every day. We also assumed that the quality of their work with teachers would be better because school-hired coaches would be more familiar with school personnel, students, and a school’s overall culture and context, which would help them build relationships with teachers and to understand the instructional strengths and needs of particular schools.

However, this is not what the data told us. Instead, the district-hired coaches reported spending substantially more time doing instructional work with teachers than did school-hired coaches. District-hired coaches spent an average of 92% of their time, averaged over two years, working with teachers in what researchers have called “potentially productive coaching activities,” such as co-teaching, modeling, observing, giving feedback, or preparing for collaborative work with teachers (Gibbons & Cobb, 2017). By contrast, school-hired coaches reported spending 40% of their time in these activities. One reason that district-hired coaches spent more time working directly with teachers on issues of instruction was that district leaders emphasized these activities as defining features of their work.

Yet, district-hired coaches’ work was not without its drawbacks: District-hired coaches tended to visit individual schools just once per week. They reported struggling to build relationships with both teachers and principals, which meant that they were disconnected from the larger instructional goals of the school and did not have equal access to all teachers. Although the situation improved if district-hired coaches worked in a particular building for more than one year, they often described difficulties earning teachers’ trust, particularly from veteran teachers, whose doors were described as “open, but just a crack.” Thus, district-hired coaches worked most often with new or struggling teachers, but principals worried that even this work was too inconsistent to support teachers in a long-term or substantial way.

Coaches’ use of time was most closely related to whether they were hired by the district or the school.

School-hired coaches had a different set of obstacles. Teachers and principals generally reported having greater trust in them, noting that they spent every day in their school and, in most cases, had worked there as a teacher prior to becoming a coach. However, their time to work directly with teachers on issues of instruction was more limited than district-hired coaches’ time because they had too many disparate jobs. School-hired coaches typically held at least one job title in addition to mathematics coach, including teacher of record, department head, substitute teacher, tutor, Title I coordinator, or building assessment coordinator. This meant that school-hired coaches spent substantial time working directly with students — tutoring, substitute teaching, or teaching courses that were created in response to low test scores. They also described spending large proportions of their time assigning and maintaining students’ log-in information for various software programs, proctoring interim assessments, collating test data, identifying students for interventions (based on their test scores), and finding curriculum for and teaching new intervention courses. Unfortunately, none of these activities helps teachers improve their instructional practice, which means that school-hired coaches did not necessarily get to make the best use of the strong relationships they built with teachers.

The role of relationships

We initially assumed that coaches who had a positive relationship with their principals would spend the most time working directly with teachers on issues of instruction and that they would report being able to work with teachers in the most in-depth and ongoing ways. Indeed, studies suggest that coaches tend to have greater access to teachers when principals publicly support or participate in their work (i.e., Gibbons, Wilhelm, & Cobb, 2017; Matsumura, Garnier, & Resnick, 2010).

However, the experience of school-hired coaches in our study complicates the picture. All of these coaches were hired by principals, and their principals reported great trust in them, describing them as knowledgeable and capable professionals who deeply understood their content. Yet, those principals assigned them many non-coaching duties, which meant that they spent less time than their district-hired peers working with teachers.

On the other hand, principals in this district were under great pressure to improve test scores, and if they did not succeed, they risked being censured in the newspaper or losing their jobs altogether. Against this backdrop, principals understandably devoted substantial resources, including the time of their instructional coaches, to collating student test data, designing intervention systems for low-performing students, and asking those whose content knowledge they trusted most (school-hired coaches) to teach or tutor students. Importantly, all 14 of the principals in our study took these steps even though they told us that their coaches’ most important job was to support teachers to improve instruction. In short, the principals were forced to make hard choices: Should they allocate coaches’ time toward activities designed to meet highly consequential, short-term goals related to students’ test scores, or toward longer-term (and more amorphous) goals related to teachers’ instructional improvement? Given this choice, most principals devoted school-hired coaches’ time toward the pressing, short-term goal.

The best of both worlds

In the second year of our qualitative analysis, the district hired coaches and assigned them to work in only one school each, effectively capitalizing on the best of both the district-hired and school-hired coaching models. Because the coaches were now accountable to district leaders — who were shielded to some extent from the accountability pressures that principals faced, giving them more freedom to invest in long-term instructional improvement — district-hired coaches were less likely to be assigned to non-coaching duties meant to help boost test scores. And because they now spent their time in a single building, they were able to develop stronger relationships with teachers and staff.

Further, in this redesigned model, principals were required to apply to the district to be assigned a coach. Specifically, they had to demonstrate their need for assistance, request a coach by name, and — most important — agree to set aside specific times for the coach to work with teachers on their instruction (which meant also that they had to reserve time for teachers to work with their coach). In short, the coaches were given a guarantee that they would be able to devote the bulk of their time to the work they were trained to do. (That year, in fact, these coaches spent 66% of their time working closely with teachers and principals, compared to 40% for their school-hired peers.)

Accountability matters

The most important implications of these findings relate to who should hire coaches and how many schools coaches should serve. While the value of teacher leadership for teachers’ professionalization and retention should not be discounted (York-Barr & Duke, 2004), our findings suggest that instructional coaches who are accountable to principals may too often be asked to attend to duties unrelated to supporting teachers’ instructional improvement, even when principals themselves highly value this goal. One reason is that principals must contend with other, competing goals, such as improving students’ test scores. When coaches are hired by the district, however, they may have greater opportunity to spend time with teachers on long-term instructional improvement goals because district leaders may be better able to shield them from extensive administrative work.

Yet district leaders must also ensure that district- hired coaches are not spread too thin across too many schools: Coaches must be allowed to work in ongoing ways in a single school, so that they might develop substantive, ongoing relationships not only with teachers but also with principals. If coaches are not in the building on a consistent basis, then principals may consider them to be marginal to the school’s improvement goals, relegating them to work with only the new or “struggling” teachers. Such teachers may very well need support, but research suggests that in those schools where teachers’ instruction improves, teachers of varying expertise work collaboratively toward a set of common goals (McLaughlin & Talbert, 2001). Besides, coaches’ expertise is too valuable — and too expensive — to be left at the margins of the important work of learning to teach for increased equity and intellectual rigor.


Bean, R.M., Draper, J.A., Hall, V., Vandermolen, J., & Zigmond, N. (2010). Coaches and coaching in Reading First schools: A reality check. The Elementary School Journal, 111 (1), 87-114.

Cobb, P., Jackson, K., Henrick, E., Smith, T.M., & the MIST team. (2018). Systems for instructional improvement: Creating coherence from the classroom to the district office. Cambridge, MA: Harvard Education Press.

Darling-Hammond, L., Wei, R.C., Andree, A., Richardson, N., & Orphanos, S. (2009). Professional learning in the learning profession: A status report on teacher development in the United States and abroad. Dallas, TX: National Staff Development Council.

Gibbons, L. & Cobb, P. (2017). Focusing on teacher learning opportunities to identify potentially productive coaching activities. Journal of Teacher Education, 68 (4), 411-425.

Gibbons, L., Wilhelm, A.G., & Cobb, P. (2017). Coordinating leadership supports for teachers’ instructional improvement. Manuscript submitted for publication.

Hawley, W.D. & Valli, L. (1999). The essentials of effective professional development: A new consensus. In L. Darling-Hammond & G. Sykes (Eds.), Teaching as the learning profession: Handbook of policy and practice (pp. 127-150). San Francisco, CA: Jossey-Bass Publishers.

Kane, B.D. (2016). Design conjectures for preservice teachers’ concept development about rigorous and equitable writing instruction. Teaching/Writing: The Journal of Writing Teacher Education, 5 (1).

Kane, B.D., Cobb, P., & Gibbons, L. (2018). Supporting ambitious instruction at scale: The role of instructional coaching. In P. Cobb, K. Jackson, E. Henrick, T.M. Smith, & the MIST team. Systems for instructional improvement: Creating coherence from the classroom to the district office. Cambridge, MA: Harvard Education Press.

Kane, B.D. & Rosenquist, B. (2018). Relationships between district and school-level policies and expectations for instructional coaching and coaches’ time use. Manuscript submitted for publication.

Knight, D.S. (2012). Assessing the cost of instructional coaching. Journal of Education Finance, 38 (1), 52-80.

Matsumura, L.C., Garnier, H.E., & Resnick, L.B. (2010). Implementing literacy coaching: The role of school social resources. Educational Evaluation and Policy Analysis, 32 (2), 249-272.

McLaughlin, M.W. & Talbert, J.E. (2001). Professional communities and the work of high school teaching. Chicago, IL: University of Chicago Press.

Sailors, M. & Shanklin, N. (2010). Growing evidence to support coaching in literacy and mathematics. The Elementary School Journal, 111 (1), 1-6.

York-Barr, J. & Duke, K. (2004). What do we know about teacher leadership? Findings from two decades of scholarship. Review of Educational Research, 74 (3), 255-316.


Originally published in April 2018 Phi Delta Kappan 99 (7), 21-25. © 2018 Phi Delta Kappa International. All rights reserved.


BRITNIE DELINGER KANE ( is assistant professor of literacy education at the Citadel’s Zucker Family School of Education in Charleston, S.C.
BROOKS ROSENQUIST is a doctoral candidate at Vanderbilt University, Peabody College, Nashville, Tenn.

No comments yet. Add Comment

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

stdClass Object ( [ID] => 68196 [display_name] => Britnie Delinger Kane [first_name] => Britnie [last_name] => Delinger Kane [user_login] => britnie-delinger-kane [user_email] => bdelingerkane@fake.fake [linked_account] => [website] => [aim] => [yahooim] => [jabber] => [description] => BRITNIE DELINGER KANE ( is assistant professor of literacy education at the Citadel’s Zucker Family School of Education in Charleston, S.C. [user_nicename] => britnie-delinger-kane [type] => guest-author ) stdClass Object ( [ID] => 68197 [display_name] => Brooks Rosenquist [first_name] => Brooks [last_name] => Rosenquist [user_login] => brooks-rosenquist [user_email] => brosenquist@fake.fake [linked_account] => [website] => [aim] => [yahooim] => [jabber] => [description] => BROOKS ROSENQUIST is a doctoral candidate at Vanderbilt University, Peabody College, Nashville, Tenn. [user_nicename] => brooks-rosenquist [type] => guest-author ) 8 |