Instructional coaching’s magical thinking problem - or 8 threats to coaching

In the last few years, donors, governments and implementers have increasingly placed their faith in coaching to help teachers transfer the fragile knowledge gleaned from a workshop into deep learning in the classroom. But how should coaching actually work?

December 08, 2020 by Mary Burns, Escola Superior de Educação de Paula Frassinetti
7 minutes read
Teacher coaching in Yemen. Credit: ECD/Mary Burns
Teacher coaching in Yemen.
Credit: Credit: EDC / Mary Burns

Perhaps no reform effort has enjoyed more rapid uptake in international education projects over the last few years than teacher coaching. For good reason. Research shows that instructional coaching, in contrast to other reforms (like smaller class size), can result in improved teaching that in turn yields more positive effects on student achievement (Kraft et al, 2018).

As a result, donors, governments and implementers have placed their faith in coaches and coaching to help teachers transfer the fragile knowledge gleaned from a workshop into deep learning in the classroom and to reform teaching.

But as this post will argue, coaching only works if it is conceptualized, designed and implemented well. And here international development coaching programs, writ large, often struggle.

The theoretical underpinnings of coaching often appear intuitive versus research based. Many coaching initiatives are grounded in incomplete theories of change; and recruitment, training and support for coaching are often inadequate.

In short, our aspirations for coaching, minus the foundations that make these hopes a reality, result in a kind of magical thinking about coaching that threatens this potentially powerful form of teacher professional development and support.

I write about these internal threats from the perspective of one who has been a school-based coach, who was trained in cognitive coaching, and whose job over the last several years has been designing and supervising coaching programs for donors in developing countries.

Threat #1: No teaching experience necessary

Coaching is highly technical. It is highly differentiated, individualized, personalized professional development that is craft focused. Teaching is extraordinarily complex—that’s why so many people struggle to do it well.

Skills demonstration; deep understanding of the many dimensions of teaching; empathy with the many travails teachers face; experience with the reality of schools, classrooms and students; offering guidance based on prior experiences; and credibility are essential coaching skills—and they are skills that come from the lived experience of having taught.

Threat #2: No assembly required

But there’s also an inversion of the above, which is, if you’ve been a teacher, ipso facto, you can be a coach! And when we do train coaches, it is often focused on filling out observation forms and donor-funded compliance and accountability forms.

But if teaching is complex, coaching may be more so. The “bricks” of strong teacher experience must be bonded with the “mortar” of coaching skills: feedback; understanding which coaching approaches to deploy (See #7); conflict resolution; change management; and communication skills like summarizing, shifting, and paraphrasing that move teachers from one performance level to another.

The best coaches make coaching look easy. But the seeming simplicity of good coaching belies the difficulty of doing it well and the range of complex and deep skills, knowledge, and dispositions that must be developed and constantly nurtured to create this expertise.

Threat #3: Coaching, mentoring…same difference

“Coaching” and “mentoring” are used synonymously in international development; and while they are similar, and both valuable, there are differences.

Coaching is a technical relationship—an intensive method of directing, instructing, empowering, and guiding teachers to achieve a particular goal or develop a set of specific skills sustained over a longer period. A coach can be an external expert or another teacher (Killion & Harrison, 2008).

Mentoring is a more developmental relationship. It is often more short term, and less technically intensive than coaching. Mentors are older, more experienced teachers, who listen, guide and support new teachers trying to “learn the ropes.” They may observe a new teacher’s class and give feedback; they may have the new teacher observe their classes and then break down the lesson and give advice.

Threat #4: The wolf in sheep’s clothing: Cascading coaching

Coaching is a repudiation of the cascade approach of capacity building. It is a recognition that we learn through sustained, ongoing practice by doing the work in our place of work. Yet how does the international development world prepare the coaches themselves? Through the cascade model. Thus, despite investments in coaching, we still can’t seem to quit the cascade.

Teaching coaches to coach via a cascade approach is dangerous for multiple reasons, two of which I’ll mention. First, it’s incompatible with the whole raison d’être of coaching. While one can cascade a discrete set of standardized behaviors, coaching involves a complex set of behavioral skills. Further, coaches need rigorous training, differentiated and ongoing support, and constant calibration and fine tuning. The cascade is not equipped to do either of these things.

Second, the cascade violates “do no harm” because it ignores the quality of the coach. A good coach can be transformative. A bad coach can be disastrous. Cascading coaching veers uncomfortably close to professional malpractice by foisting a lot of unprepared coaches onto a lot of vulnerable teachers.

Threat #5: Coaching = observation and feedback

Observation and feedback are certainly the core of coaching (Joyce & Showers, 1981). But these two elements were envisioned as part of, not the entirety of, coaching.

This paradigm is the most common feature of donor-funded projects that I have helped supervise over the last few years. Indeed, it often is the coaching intervention. This paradigm needs serious re-examination for three reasons.

First, it rests on a weak theory of change—the idea if we tell someone how to do better, they’ll do it. Yet, research is clear that verbal instruction is often “inadequate,” (Gladwell, 2005); that our first impulse in receiving information that points out our shortcomings is to reject this information; and that attempting to promote excellence by focusing on failure is futile (Buckingham & Goodall, 2019).

This is why well-trained coaches are critical to observation-feedback. A cognitive coaching approach, for example, helps the teacher become more “coachable” by skillfully guiding the teacher to evaluate his/her own work and reducing resistance to change (Garmston et al, 1993).

Second, “coaching = observation and feedback” strips away the critical instructional supports that make coaching so potentially transformative in the first place. Research on change shows that we learn best through models and direct experience (Gladwell, 2005).

As an example, in Indonesia from 2008-2010, EDC used a “gradual release” coaching approach where highly trained coaches helped teachers set instructional goals, modelled a lesson that incorporated these goals, helped teachers design a similar lesson, co-taught the lesson with the teacher, observed and provided feedback as the teacher taught a similar lesson alone.

This continuous cycle of goal setting, modeling excellence, reinforcing and refining teacher skills to follow these models, and focusing on teacher strengths and agency, helped teachers develop the confidence, volition and skills to measurably improve practice (Ho & Burns, 2010).

The third reason is below.

Teacher coaching in Yemen. Credit: EDC / Mary Burns
Teacher coaching in Yemen
Credit: EDC / Mary Burns

Threat #6: Coaching = evaluating

Coaching is a supportive relationship, not an evaluative one. Yet many coaching programs confound classroom observations with accountability. Though coaching can be directive, coaches should not be supervisors. Though monitoring is part of coaching, it should be for purposes of improvement, not compliance. And though assessment is part of coaching, it should be formative rather than summative.

Coaches observing teachers is about a growth mindset and improvement—not accountability.

This conflation muddies the coach-teacher relationship because it undermines the trust and support that are intrinsic to that relationship.

Threat #7: Coaching is non-differentiated

In fact, coaching is highly differentiated. There are many different types of coaches; coaches have multiple roles; and there are many “levels” of coaching—from light-touch to EDC’s intensive approach mentioned above.

There are also numerous coaching approaches. Broadly, coaching can be categorized as “directive”—informational, technical, prescriptive coaching versus “facilitative” coaching that is supportive and collaborative.

The kind of approach that a coach uses, and the frequency and titration, depend on the type of professional development, desired outcomes for teachers, and the complexity of the intervention.

For example, if the goal is to help teachers correctly utilize a series of defined reading approaches, then a directive approach may be more appropriate. If the goal is to help teachers think strategically about a problem or enhance team building, then a facilitative approach may be more appropriate. In reality, coaches frequently alternate between these approaches.

Threat #8: Coaching is too expensive

Finally, there is no question that coaching is resource intensive (as are centralized workshops). However, the widespread belief that it is “too expensive” for international development programs dwells more in the realm of supposition than fact. We have minimal data on the true cost of coaching, and we lack frameworks that measure its costs vis-à-vis other types of teacher professional development.

Coaching may in fact be “too expensive,” (and not cost effective) but until such claims are substantiated by evidence, we should not pre-emptively dismiss coaching as financially unfeasible.

And we can cut down on the costs associated with coaching— for instance, employing peer coaching, group coaching and hybrid and technology-based coaching. The latter has yielded mixed results, as I’ll discuss in the next post on this topic.

Read the other blogs in this series:



  • Buckingham, M. & Goodall, A. (2019-April-May). The feedback fallacy. Harvard Business Review, 92-101. Brighton, MA: Harvard Business Publishing
  • Garmston, R., Linder, C., & Whitaker, J. (1993). Reflections on cognitive coaching. Educational Leadership, 51(2), 57–61.
  • Gladwell, M. (2005). Blink: The Power of Thinking Without Thinking. New York, NY: Back Bay Books.
  • Ho, J. & Burns, M. (2010). Evaluation of Pilot 1 Decentralized Basic Education 2 (DBE 2). Unpublished evaluation report submitted to USAID.
  • Joyce, B. R., & Showers, B. (1981). Transfer of training: The contribution of "coaching.” Journal of Education, 163-172.
  • Killion, J., & Harrison, C. (2008). Taking the lead: New roles for teachers and school-based coaches. Oxford, OH: National Staff Development Council.
  • Kraft, M.A., Blazar D, Hogan D. (2018). The effect of teacher coaching on instruction and achievement: A Meta-analysis of the causal evidence. Review of Educational Research, 88 (4):547-588
Quality teaching

Related blogs


Mary, thank you for this. These are spot on and merit to be acknowledged by our sector! Curious, in lieu of coaching, what TPD and teacher support mechanisms have you found to be most effective and doable at scale with limited funding since let's face it, our targets are often quite high but the cost per teacher is often so low when comparing to what we spend in the US on teacher PD and support.

Hi Nathalie, Sorry for the delay in responding.

There’s not great research on TPD effectiveness, particularly models of PD, in developing country contexts so we don’t know what works best or what is most cost effective. I think the reason we can’t scale in international development is because we focus on “spread” (pushing out an innovation to as many people as possible) -- not “depth” (making sure that teachers have real mastery of the innovation itself). We still don’t know how to scale quality in our field.

• The best examples of capacity building at scale I've seen: The Weekend School, a Dutch initiative that you can read about here: The woman who started it, and still runs it, essentially franchises the whole operation so that there are now dozens of Weekend Schools in Europe (mainly Belgium and NL) It’s run by young people for young people but the Weekend School spends YEARS preparing, grooming and building the capacity of its members—and they have to pass through several stages. This is the only example of successfully scaling quality I’ve ever seen in education (I did some research on these schools in 2016-2017).

• In terms of PD and scale, I don't know if there's one model of PD that is most scalable. I think what is scalable goes back to 1) what teachers are supposed to learn (simple vs. complex; knowledge vs. behavior) 2) existing mechanisms for diffusing these innovations and 3) capacity in the system to help teachers learn whatever this is? If those things exist, then the PD becomes -- which format is best for learning X (a study group? A workshop?) vs which is most scalable.

In terms of scale, this is where technology should/can/is supposed to help. You can add a teacher at marginal cost in online learning and you can reach more teachers through several sections of the same online course. You can provide blended coaching and support to teachers via web conferencing, Skype calls, phone calls, texting. But you can’t do any of this is there’s no Internet and if teachers don’t have the needed technology and skills to use the technology. So the technology that best allows us in theory to scale all kinds of PD (Web based) is inaccessible to many/most of the teachers who would benefit from it.

I think highly structured technology (with a broadcast element) like radio (IRI/IAI) and TV have shown themselves to be very scalable while maintaining fidelity and assuring quality; mobile phones (texting and calling and simple supports) less so but okay. I know RTI uses Tangerine for coaching at scale, but I’m not familiar with it. It would be worth talking to RTI.

I'm wondering if I answered every question EXCEPT what you asked...:)

Thanks for reaching out. Merry Christmas!

In reply to by Mary Burns

These are great! As always, I appreciate your insight.

Apologies for not thanking you earlier-I didn't get any notification you'd replied :)

Leave a comment

Your email address will not be published. All fields are required.

The content of this field is kept private and will not be shown publicly.

Plain text

  • Global and entity tokens are replaced with their values. Browse available tokens.
  • No HTML tags allowed.
  • Lines and paragraphs break automatically.
  • Web page addresses and email addresses turn into links automatically.