Does Coaching Work?

Generally, the phrase “what works” means boosting test scores in low-performing schools beyond business-as-usual. The short answer is that coaching sort of does work (see e.g., Biancarosa, Bryk, & Dexter, 2010). Several studies, of varied quality, have found that coaching is present in programs that produce achievement gains in randomized controlled trials (RCTs). A meta-analysis of 23 studies of programs with coaching as a component (Kraft, Blazer, & Hogan, 2018) demonstrated an average effect size of 0.15 on achievement. Biancarosa and colleagues (2010), by contrast, reported an RCT designed to isolate the extent to which coaching per se caused changes in student achievement. They, too, reported modest effect sizes for some of the measures in their study.

There’s a theory of action at work here. Professional development (PD) is more than coaching... and less than coaching. PD aims to improve instructional practice, and the improved instructional practice is supposed to (especially if the practice is “research-based”) improve student achievement. So where does coaching fit in?

In the pre-coaching days, PD took place as didactic (sit-n-git) events: lectures and workshops that described and explained things. Transfer of any learning that ensued was pretty much up to the educator (most often a teacher, and more rarely a principal). So not much transfer took place.

The purpose of coaching (as first articulated by Joyce and Showers in 1982) was to improve the transfer of learning from (or in) PD. In other words, coaching isn’t a stand-alone practice. It’s embedded in a larger picture.

Teachers receive coaching, but student achievement is somewhat remote from the direct aim of coaching: to help teachers (and principals) change their practices for the better. Does coaching actually change practice (adult behavior) for the better? As you might expect, the answer here is clearer. When coaching is done well, and especially when it’s part of a well-conceived and responsibly implemented reform effort, professional practice does change (Kraft et al., 2018; Pomerantz & Pierce, 2019). That’s the good news.

The bad news? Good coaching within a well-conceived and well-implemented reform effort is understandably difficult. And coaching efforts alone fall short of expectations (e.g., Deussen, Coskie, Robinson, & Autio, 2007).

So what? So the more nuanced answer is that coaching, in typical practice, too often fails to accomplish what it otherwise might. When an organization doesn’t pay PD the attention it deserves, and is neglectful of the coaching component within the PD, coaching can’t work. Neither, of course, would anything else under such circumstances. Just adding coaching positions, without a robust PD context, is a common misstep (Mangin & Dunsmore, 2015). Additionally, Kraft and Blazer (2018) conclude that, even within a successful PD effort, simply adding coaching positions presents problems. Coaching is labor intensive, and cultivating the capacity to provide good coaching is an on-going challenge.