Unbalanced Literacy

(Note: A PDF of this post can be downloaded here.

Over the last year or so, an education reporter named Emily Hanford has published a series of exceedingly important articles about the state of phonics instruction (or rather the lack thereof) in American schools. The most in-depth piece appeared on the American Public Media project website, but what are effectively condensed versions of it have also run on NPR and the NY Times op-ed page.

If you have any interest in how reading gets taught, I highly recommend taking the time for the full-length piece in APM: it’s eye-opening and fairly disquieting. While it reiterates a number of important findings regarding the importance of phonics, its originality lies in the fact that Hanford takes on the uneasy truce between phonics and whole language that supposedly put an end to the reading wars of the 1980s and ‘90s, and points out that so-called “balanced literacy” programs often exist in name only.

In principle, this approach recognizes that both development of sound-letter relationships and consistent exposure to high-quality literature are necessary ingredients in helping students become proficient readers. What Hanford does, however, is expose just how vast a chasm exists between theory and reality. In many schools, phonics is largely neglected, or even ignored entirely, while discredited and ineffective whole-language approaches continue to dominate.

To be clear, “reading” in the complete sense of the term is an incredibly complex, multifaceted act, one that draws on the ability to form letter/sound relationships, recognition of sight words, vocabulary, syntax, and background knowledge, among other things. “Decoding,” in contrast, refers to the act of being literally able to translate squiggles on a page into words (although confusingly, it’s often referred to as “reading” as well). It’s a key component of reading but obviously in no way a substitute for the full range of skills required. What Hanford is primarily concerned with is the latter, and neither she nor any of the experts she cites view phonics as more than a single piece—albeit an extremely fundamental piece—of the reading puzzle.

So, it is all very well and good to agree that students need to learn some phonics, but what exactly does “some” mean? 10 minutes a day? An hour once week? An hour a month? Once a year? Hanford recounts the story of a parent who, concerned about her child’s reading, asked the teacher when the class would cover phonics. The teacher responded that she had covered phonics but that the child had been absent that day.

And then there’s this story, which I think speaks for itself:

[Bethlehem, PA, district literacy supervisor Kim] Harper went to a professional development day at one of the district’s lowest-performing elementary schools. The teachers were talking about how kids should attack words in a story. When a child came to a word he didn’t know, the teacher would tell him to look at the picture and guess. The most important thing was for the child to understand the meaning of the story, not the exact words on the page. So, if a kid came to the word “horse” and said “house,” the teacher would say it’s wrong. But, Harper said, “if the kid said ‘pony,’ it’d be right because pony and horse mean the same thing.”

Before I go any further, let me explain that I find Hanford’s work interesting for a few reasons. First, although pretty much all of my work has been with high school and college students, what happens in elementary-school classrooms doesn’t stay in elementary school classrooms; the consequences of what students do or don’t learn when they’re six can have far-reaching implications.

When I started tutoring, I was immediately struck by the fact that when some of my students encountered an unfamiliar word, they made no attempt whatsoever to sound it out; the prospect literally didn’t seem to occur to them. Rather, they’d read the beginning of the word and guess, or just substitute a similar-looking word entirely. Furthermore, they had a tendency to do things like skip entire lines of text without noticing, and their eyes often seemed to dart around the page randomly. Some of these students were clearly bright, and they didn’t have diagnosed learning disabilities either; it just seemed that the conceptual tether between letters, sounds, and words had never been properly developed. I had never encountered anything like it, and I was both stunned and fascinated. I wanted to understand what sort of system could possibly produce reading problems so bizarre. I also wanted to understand how these students—who were not exactly disadvantaged or at risk of flunking out of school—had managed to make it to eleventh grade without any intervention.

The second reason is that in taking on the gap between what balanced literacy is supposed to be vs. how it actually plays out, Hanford calls attention to major shortcomings in the way teachers are trained, and about how broad policy decisions get turned into actual curriculum. It also raises serious questions about the default American tendency to turn pedagogy into a matter of individual preference, without consideration of vast differences in knowledge and preparation among teachers. It is one thing to train teachers well and then give them lots of autonomy in the classroom, but it is another issue altogether when teachers are trained poorly and then left to their own devices.

The third reason—the one I personally find most interesting—is that the debate over how reading is taught can actually be seen as a reflection of a larger cultural crisis over the role of expertise. The various players involved cannot even agree on what sort of evidence should be considered credible, or on who can be considered an authority, or on what constitutes proof of effectiveness. One of the most disturbing quotes in Hanford’s full-length article comes from a reading teacher in Mississippi who, when presented with the evidence in favor of teaching phonics, waves it away with a dismissive, “Well, that’s your science” (emphasis mine). That pretty much sums things up.

As in all things involving education, the devil is in the details. To be taught effectively, reading, like any complex task, must for most students be broken down into clear, logically and coherently sequenced steps. The problem is that teachers who are themselves the product of an incoherent, poorly sequenced, gap-ridden, and haphazard curriculum often have literally no experience with or basis for developing this kind of systematic approach. This is further compounded by the fact that many teachers are trained in a system steeped in airy notions about natural learning and are inclined to disregard conflicting ideas. Indeed, when someone whose default model is chaos first encounters a structured curriculum, their instinct may be to view it as overly rigid and nitpicky, and to reject it out of hand. So even when some phonics does get taught, it may still be far less structured than what some students require.

E.D. Hirsch has written about extensively about the romantic belief that learning should be “easy” and “natural,” a belief that has become one of the great guiding principles of contemporary American education. Not only is this belief often at odds with reality, but adherence to it is often expressed in a stubborn refusal to acknowledge the difference between how novices and experts engage in a particular skill. (Interestingly, this willful misunderstanding tends to dissipate when non-academic activities, particularly sports, are involved.) The result is a confusion between ends and means, and a rejection of the possibility that intermediate steps might be required for mastery. Because practicing a skill in finished form feels natural to experts, they are inclined to forget how much practice went into their own acquisition of it and believe that it is unnecessary to break things down—or worse, that doing so is counterproductive.

Hanford, for example, interviewed Mary Ariail, the former chair of the Department of Curriculum, Instruction and Special Education at the University of Southern Mississippi, who “believes kids will be distracted from understanding the meaning of what they’re reading if teachers focus too much on how words are made up of letters. “One of the ideas behind whole language is that when [reading] is meaningful, it’s easy,” she said. “And when it’s broken down into little parts, it makes it harder.” The operative word here is “believe.” One thing I find truly baffling is how students are supposed to read “meaningfully” when they literally cannot figure out what the words say. Are they supposed to memorize every new word and treat English like a series of pictographs, akin to Chinese? English spelling may be irregular, but the myriad exceptions do not negate the fact that rules do exist, and many words can be figured out through a logical process.

Hanford also interviewed Candy Maldonado and Adrienne Ibarra, two teachers from the Bethlehem, PA, a formerly low-performing district that saw enormous gains after implementing a highly structured district-wide phonics program. In the past, the teachers “had no idea how kids actually learned to read. ‘It was just that they do,’ Ibarra said. ‘Almost like it’s automatic,’ Maldonado added.

As scholars such as Mark Seidenberg, a cognitive neuroscientist at the University of Wisconsin and the author of Language at the Speed of Sight, have determined, however, this is not in fact the case: reading is a relatively recent development in human history and draws on different parts of the brain than does a truly natural skill like speech. With the exception of a very small percentage of “natural” readers, children must be explicitly taught the relationship between letters/combinations of letters and sounds in order to learn to decode.

I think, however, that there’s yet another dimension to this problem, namely that scientists like Seidenberg and some proponents of whole language do not even necessarily agree on what learning looks like, or indeed what it is. I think that this is part of the reason that cognitive science and education people often end up talking past each other, but it rarely gets discussed. Hanford, for example, recounts the following from her interview with Stacey Reeves, an education professor at the University of Southern Mississippi:

In the early 1990s, before she started her Ph.D., she was an elementary school teacher. Her students did phonics worksheets and then got little books called decodable readers that contained words with the letter patterns they’d been practicing. She said the books were boring and repetitive. “But as soon as I sat down with my first-graders and read a book, like ‘Frog and Toad Are Friends,’ they were instantly engaged in the story,” she said. She ditched the phonics workbooks and the decodable readers. “And once I started teaching in a more whole way, a more encompassing way of the whole child — What does this child need? What does that child need? Let’s read more real books,” she explained, “my teaching improved, the students learned more. I feel they came out the other side much better.” She admitted she had no evidence her students were learning more, but she said they seemed more engaged.

If a teacher has been trained to believe that children can only be learning if they are smiling and appear excited, she will be inclined to avoid activities challenging activities that produce evident discomfort. Think of a child trying to sound out his or her first words, face screwed up in concentration, stammering them out syllable by syllable, and contrast that image with one of a child easily and proudly “reading” a text that he or she has mostly memorized, or listening intently to an engaging story. The former might be an important step towards gaining proficiency, but it sure doesn’t look like much fun. In fact, it seems downright unpleasant. Besides, children sounding out their first words are typically working so hard that they don’t have mental bandwidth to enjoy, or even think about, the meaning. So if the true purpose of education is to inspire joy and confidence, why not just emphasize activities like the latter? For teachers in this camp, students’ visible enjoyment of an activity is itself the only proof of its effectiveness one needs. Being excited about learning is seen as equivalent to learning.

As the cognitive scientist and reading specialist Marilyn Jäger Adams has pointed out, much of education is essentially a con job that involves persuading students that they really do want to learn whatever it is you’re trying to teach them. This is a huge part of what makes teaching difficult! But Reeves, having observed that the drills were boring, made no attempt to understand why they might have been created, or to make them more interesting, perhaps by turning them into a game. She merely abandoned the activity entirely in favor of something that provided more immediate gratification for both her and her students. (Quick personal anecdote: one of my favorite activities in seventh-grade French class was verb “races,” in which two students competed to see who could conjugate a verb more quickly. I thought they were loads of fun, and boy did I get good at conjugating.)

It is also not implausible that Reeves might have projected some of her own distaste for drills onto her students. Children, unlike adults, tend to enjoy repetition—think of a four-year-old who demands to be read the same story night after night. What happens when adults cannot distinguish between their own preference for variety and children’s actual needs?

Compounding the problem is pressure from parents and administrators to keep kids entertained, not to mention technology that trains children to expect instant gratification. Even if teachers want students to engage in challenging activities that require significant focus, students may lack the attention span to participate fully.

But the desire to avoid even moderate unpleasantness for the sake of long-term rewards, can create real problems down the road. (Incidentally, Adams points out that the amount of difficulty a student experiences in learning to read does not necessarily affect their ultimate level of proficiency.) A student who picks up just enough to slide through during elementary school might encounter real difficulties later on—ones that often require a lot of time and money to remedy.

Teachers only see what goes on in their classrooms; they may be entirely unaware of the amount of support some of their students receive on the outside. Indeed, the comments section accompanying Hanford’s Times piece was filled with anecdotes from parents who either taught their children phonics on their own or paid a tutor to do so. To return to one of my favorite themes, the problem doesn’t go away; it is merely privatized. Phonics-based reading centers exist in spades; when run by well-trained professionals, they’re incredibly effective, and so people are willing to pay. Private learning centers that specialize in whole language… well, they really don’t exist. The result is that students whose families can’t afford alternative sources of education often just don’t learn. It also doesn’t help that the people who come down most strongly in favor of “natural” learning tend to be fairly privileged. (The letters to the editor in response to Hanford’s Time’s piece, for example, came down solidly in favor of whole language; if that’s not broadly representative of Times readership, then it at least makes clear where the editorial board stands.) In contrast, the parents of children who stand to gain most from intensive phonics instruction are more likely to be poor and to lack the knowledge/time/energy/savvy to lobby for it.

How, though, does one cross the barrier between serious researchers like Seidenberg, who are effectively banging their heads against the wall in an attempt to get their findings translated into policy on a broad scale, and a cadre of educators increasingly wary about outside intrusion?

One of the side effects of the test-and-punish “reform” regime that has dominated since the implementation of No Child Left Behind is the backlash against the validity of testing in general. The fact that many of the standardized (Common Core-aligned) state tests were poorly designed and unfairly used for punishment has led to a tendency among some teachers to dismiss the idea that scores might in some cases be indicative of serious underlying problems, and to view any attempt at change as part of an technocratic master plot to reduce education to a series of easily plottable data points. Peter Greene, who writes the Curmudgucation blog and with whom I often agree, takes up this argument in a way I find worrisome:

Yes, we have tests. But testing and pedagogy of reading are mostly locked in a tautological embrace. I think decoding is The Thing, so I create a test that focuses on decoding, then implement classroom practices to improve decoding skills and voila–I scientifically prove that my decoding-based pedagogy works. Mostly what we’re busy proving is that particular sorts of practices prepare students for particular sorts of tests. Big whoop.

With all due respect, teaching phonics is not about “preparing students for particular sorts of tests.” That’s a very cynical—and unfair—minimization of the issue here. The goal of people like Hanford and Seidenberg is not to claim that phonics is some sort of magic bullet for making students into expert readers (Seidenberg: “Phonics is just one specific component of learning to read that’s important at a particular point in a child’s development”) and then design a series of exams to “prove” that that is the case, or use it as an excuse for education companies to develop and market expensive new curriculums. This is not even fundamentally about testing, except in the general sense that difficulties in decoding are inevitably linked to difficulties in comprehension (you can’t understand what you’re reading unless you can figure out what the words literally are), which in turn are expressed in low tests scores. And although scores are of course limited in what they can reveal, the finding 2017 NAEP finding that 32% of fourth-graders are reading below a basic level suggests that something is amiss.

However, if the assumption is that low test scores are only an excuse for punishing teachers or implementing more market-based solutions, then it’s easy to dismiss poor results as meaningless. Where things get murky is that the premise for this belief is not entirely unfounded—some of the tests are very poorly designed and developmentally inappropriate, and in recent years, scores have been used to evaluate teachers in all sorts of wildly unfair and arbitrary ways. At the same time, though, many teachers—no matter how dedicated or well-intentioned—are not trained to teach reading effectively, and many children have unnecessary reading problems as a result. All of those things can be true simultaneously. The dysfunction exists at multiple levels.

A big part of the problem, I think, is that the concept of expertise in education has been so thoroughly muddied. Over the past decade, well-intentioned but clueless billionaires (Gates, Zuckerberg) and their cronies (David Coleman) have bumbled into the education system, convinced that their disruptions and innovations would magically fix American schools. Despite their lack of knowledge, or perhaps because of it, they have blithely presented themselves as authorities, and they have the money and power to inflict enormous amounts of damage. After being subjected to their constant and ineffectual reforms, as well as their condescension, it’s understandable that teachers would become defensive when confronted with perceived attacks from outsiders and want to protect their students from yet more turmoil.

The problem is that Ph.D.s in cognitive science like Seidenberg and UVA’s Daniel Willingham are in a fundamentally different category from the fake “experts” wreaking havoc on the American educational system. They may not work with young children in a classroom on a daily basis, but they’ve spent decades studying what reading is, how we learn to do it, and what can potentially go wrong. To lump them together with slick, ignorant, tech-driven “reformers” is unfair and ignorant, and it allows the type of approach touted by Reeves to continue unchecked. It also contributes to the idea that science is nothing more than a matter of opinion, or a partisan tool. And to make things even worse, some of these researchers’ work is cited by defenders of programs such as Common Core, tainting it by association.

As Seidenberg points out, though, “in order to grasp the research, [teachers] need basic scientific literacy to be able to understand it. They can dismiss [what I’m saying] or they can share my outrage.” If teachers lack the requisite scientific background and are uninterested in or downright averse to obtaining it, it becomes impossible to find common ground. This is also where a form of the Dunning-Kruger effect comes into play: the teachers who have the least knowledge of the science and the deepest attachment to ineffective methods will also be the ones most likely to reject alternate approaches. There is no better way to incur the wrath of ignorant people than to make them feel they are being talked down to. (They may, for example, perceive the use of even simple scientific terminology with which they are unfamiliar as impossibly technical and intolerably condescending.) This is hardly to imply that scientists know everything—they would be the first to admit that’s far from the case—but they do know some very important things that deserve to be taken seriously.

Given how toxic these debates are, the desire to throw up one’s hands and back away, announcing the whole debate is ridiculous and a waste of time, that of course students need phonics as well as exposure to authentic texts, and hey, wasn’t this whole thing settled a long time ago, and shouldn’t individual teachers get to decide what’s right for their students? But that leaves us right back at Hanford’s point: the debate may have been settled in theory, but it is nowhere near settled in practice—and ultimately that is what counts. Although there has been some improvement, at present only 37% elementary and special education programs appear to be teaching phonics. At some point, it becomes necessary to go beyond the platitude of “you know, we all really do agree here” and look at the nitpicky details of who is actually teaching what, and when, and how. These are things that Americans are notably averse to (people who are deeply invested in policy details are predictably derided as nerds, wonks, drones, etc.); it’s a very deeply ingrained attitude, and it won’t be easy to overcome. It also means discarding the “every child learns in a totally unique way” dogma and recognizing that most children can in fact be taught to decode competently through an established sequence of steps.

For me, the question is this: at what does point the evidence for a particular approach become sufficiently convincing that it seems more necessary to impose some sort of top-down implementation than to resort to the default and just let people do their thing? And if such a system is implemented, how it possible to ensure that the people in charge actually know what they’re talking about? Both Seidenberg and Daniel Willingham have made the point that if change is to occur on a broad scale, it will necessitate revamping the teacher-education process and creating closer links between science and education faculty so that aspiring teachers actually gain the necessary skills and are exposed to the science—what it does know as well as what it doesn’t—as part of their training. Given that professors of cognitive science and professors of education currently exist in separate orbits, that’s probably not going to happen in the near future, but in the long term, it’s something to work towards.

And if that means reigniting the reading wars, then bring them on.