Have We Ever Done A Good Job Teaching Writing?

And Other Questions That AI Is Inspiring People to Ask Me

A Book, titled “Forget Me Nots” with a Hand-Drawn Cover, authored by my daughter Lillian Hart-Davidson when she was in 2nd grade

The rather sudden emergence on the scene last November of ChatGPT, the Generative AI chatbot that drafts impressive simulations of human-created texts, has people interested in talking to writing experts. This is unlike any other point in my now 25+ years studying writing and composition, I should add, and that period includes the advent of the internet, the world-wide web, and mobile phones to name a few other transformative communication technologies you may have heard about.

For how long folks will stay interested, I do not know, but I’m glad that folks are asking questions about writing and learning to write! Here’s a question a reporter asked me recently:

“For this story, I’m less interested in why writing matters, and more interested in what the evidence shows about whether public schools have ever, collectively speaking, done a good job of teaching it, particularly in Title I schools.

Can you please tell me, in a nutshell, what you know about that particular question and say a bit about how you might address it?”

Sure thing. Here goes…

Practice. Deliberate Practice. That’s How Writers Learn

It’s a good question and I’m happy you are asking it. Here’s how I would answer: I think that we have had good evidence since at least 2007 with the publication of: The Nation’s Report Card: Writing 2007 about what the most effective interventions — activities we ask students to do — that drive learning in writing are.

Perhaps the most straightforward account of these is presented in this 2009 article by Kellogg and Whiteford titled : Training Advanced Writing Skills: The Case for Deliberate Practice.

In a nutshell, what helps students improve is a program of deliberate practice with an emphasis on the following four activities:

1. Giving high quality (criterion referenced, specific, and actionable) feedback on others’ writing
2. Planning Revision focused on higher order concerns (not merely grammar, spelling, etc)
3. Criterion-referenced review (a.k.a. critical reading)
4. Reflective writing about all of the above, referencing specific learning goals related to writing

Now, these are not individually new things. On the contrary, they are all common elements of a process approach to writing pedagogy.

But what is new and what became more clear in the late 2000’s is that these are difference makers, pedagogically speaking. The evidence for that is perhaps most succinctly concentrated in this metanalysis by Graham and Perrin: A meta-analysis of writing instruction for adolescent students.

Since that study came out, other researchers have set about examining these difference-making interventions in more detail. The evidence for the effectiveness of these four interventions has only gotten stronger. Here’s a good example from Cho and Cho: Peer Reviewers Learn From Giving Comments.

But If We Know What Works, What Gives?

Since you asked about Title I schools (these are schools that qualify for federal aid because they are helping low income students) it is also important to point out that achievement in writing is very stable over time and shows a gap. Here’s a study that helped to establish that gap that was published around the same time period as the other pieces I have mentioned: The State of Writing Instruction in U.S. Schools: What the Data Tells Us.

If we know what works (or can work) to improve writing, why do we still see that achievement rates are essentially stagnant? As a researcher, I must say that we do not know. But I do not think it is down to one thing. I think it is a few things.

One is that teaching is not like medicine where we have a culture of evidence-based practice. Where we select interventions based on the evidence of which activities work the best. If we did, we would see much more of a focus on deliberate practice, and on how much and what kinds practice students are doing, as the way to improve learning. Because that is what the evidence shows. We, as a culture, have invested instead in summative and standardized assessment of outcomes with almost no attention to what students have done that would predict good outcomes.

That is…madness to me. It’s like buying better and better thermometers to take temperatures with while not doing anything shown systematically to reduce a fever.

Another issue is that while we have made some additional progress since 2009 or so, we still have too few indications of what “deliberate” practice looks like that makes a difference in performance for specific groups of students. How many times do they need to practice? with what frequency? with what level of difficulty for each rep? etc. These are things my own research has looked at, and the fact that I am able to name these qualities of practice as things we can enhance or moderate to see different results is not (yet?) a common way of thinking about teaching writing for all teachers.

Writing is still taught more like the country doctor seeing patients than with an evidence-based approach, if you pardon the overly clinical metaphor. Each individual teacher does what they think is best but in reality there is very little awareness of what practice routines in writing instruction are like and how well students practice over time.

We have at the broad level across the K-16 learner’s journey, after all of this time focused on the writing process, very little understanding of What Writers Do (n.b. Gallagher’s piece is highly recommended reading on this topic).

Finally, I would say that achievement gaps most likely represent structural barriers to ensuring that students can​ practice the things I list above with adequate frequency, intensity, and quality in many schools.

It takes more than just knowledge, care and persistence from a dedicated teacher to create a classroom space where that kind of practice can consistently happen. It also takes an administrative environment that focuses less on summative outcomes and more on formative feedback to ensure that students are practicing well enough to improve. The incentives and pressures on teachers are all stacked against that in many, if not most educational situations.

Making This Case and Trying To Make Things Easier for Teachers and Learners…

I’ve spent a big chunk of my career on it! I’ve made this argument, with similar evidence, many times. Enough to know what strikes people as making a lot of sense (practice!) and what takes them by surprise (you have to let THEM give the comments because that’s the best practice). I also know what folks say the obstacles to teaching like this are. They are formidable but tractable.

The hardest thing for folks is spending less time on three things that don’t help much at all when it comes to improvement in writing: 1) in class discussion of readings, 2) drafting, 3) teachers giving feedback.

I likely lost some of my colleagues just now, I know. I’ve seen it. But the first two things take up much more time in the curriculum than they should in many instances, leaving too little time for review and revision. As a result students don’t get enough practice repetitions in the things that really matter.

And when the teacher is the main source of feedback, well the teacher is learning a lot. But the students are not learning as much. Because its the givers who gain most from the practice of framing high-quality feedback. What is that, you may ask? Comments that 1) are easily understood by the receiver, 2) reference shared criteria, and 3) are actionable, allowing the receiver to make a change to their writing that will improve it.

Giving helpful feedback is the most powerful thing you can ask students to do. It is the high-intensity interval training of writing instruction.

And this is why I have devoted a decade and a half of time creating online spaces that make it easier for students to practice these things. More criterion-referenced review and revision, more often, and with consistency and quality checks built in. I teach this way based on the evidence I cited above and more, including my own two plus decades of experience. And I want it to be easier for others to teach this way too.