Why do unproductive teaching practices persist?

A must-read blog by Esther Quintero

Esther Quintero of the Shanker Institute has written another insightful blog that should be read by anyone who cares about teachers, learners, readers, and reading. She asks, and I think provides an answer to, why so many teachers assign too much importance to context for teaching students to read and to recognize words.

I suggest you stop reading this and go straight to her blog. But fwiw, here are some of my thoughts.

Prominent use of context—pictures, syntax, asking “what would make sense?”—is of course a key feature of “three-cueing,” “balanced literacy,” and Lucy Calkins’ three-cueing-by-any-other-name, “MSV” (meaning, syntax, visual). What is decidedly not a feature of any of these is the critical role letters and their corresponding sounds (technically known as grapheme-phoneme correspondences, or GPCs for short) play in successful reading and word recognition.

Although there is much more to reading than letters and sounds, they are central, even non‑negotiable, to reading and recognizing words. Reading and recognizing printed words accurately, automatically, and efficiently—not simply decoding—are essential for successful reading development. I won’t belabor the point, since I’ve addressed it in other posts on this Substack. Feel free to email me if you can’t find them.

One of the challenges for teachers is that while learning to read requires learning to use GPCs, not all learners need the same amount or intensity of instruction to achieve what is necessary to enable reading to “take off.” Mark Seidenberg calls this take-off point reaching “escape velocity.” David Share refers to it as the point at which learners begin to “self-teach,” able to make continued progress with minimal, if any, additional grapheme–phoneme instruction.

Some learners need a great deal of this instruction, most need less but still some (and variable) amounts, and still others learn to read so easily it seems intuitive to them, as if it doesn’t much matter whether teachers use grapheme–phoneme instruction (aka phonics) or context or some combination. Understandably, teachers can assume there are many ways to learn to read, and “phonics instruction” is just one in a menu of options.

But the fact is there’s strong consensus among those who know and understand the current state of reading research that teaching students to map symbols to sounds—with context playing a (dis)confirmatory role—is a much better way to teach learners to read, identify, and recognize printed words.

But, again, up to the point where such instruction is no longer, or just minimally, necessary.

And here’s the rub: It is not trivial knowing how much phoneme-grapheme instruction, in relation to instruction in language, vocabulary, knowledge, and other important skills, is necessary for any child or group of children. Gaining this information for your students introduces additional complexity to an already complex teaching task.

As the late great reading researcher Carol Connor demonstrated before her untimely passing, too much phonics instruction can truncate reading growth by crowding out attention to language, vocabulary, and knowledge. On the other hand, not enough phonics instruction and practice reading will do so as well although for different reasons.

So why do teachers persist in using a practice that is discredited in the research literature? Esther suggests it’s because context initially helps students get the words right—an insight supported by research she cites. This, then, creates the impression that using context works better for some (maybe most or even all) students, which then further cements the assumption that context and meaning-first is a valid way to teach reading.

The problem, according to the study Esther draws on, is that the immediate advantage of using context soon dissipates.

Students in the study were presented with words they were unable to read correctly (but were at their reading ability level according to prior assessment) and were instructed to try to read the words in one of two conditions.

The first was in isolation, where students were presented only with the individual target words and could therefore only use letters and sounds to read or attempt to read them. The second was in context, where students heard a short paragraph read aloud as they followed along, up to the target word they were to attempt to read. In this condition, they could use context, letters and sounds in the word, or both. In the other condition, only letters in the words and their corresponding sounds were available. They were the only “cues,” so to speak.

Students in the context condition read more words correctly when asked to try and read the words in one of the two conditions. This replicated previous findings.

However, there was a reversal of fortune a week later when longer term retention of word reading was assessed.

Compared to the words students learned in context, the words read in isolation were nearly 50% more likely to be recognized when the students saw them in print again. The researchers called this a “A paradoxical relation between reading and learning,” because there was greater short-term successful word reading in context, but it came at the expense of retention a week later.

Esther draws several important lessons from her astute analysis about teaching reading and also about taking seriously teachers’ perceptions about what works and why. I encourage everyone to go to her blog, which you can access here in case you’re still reading this.