Some instructional moves are so common that almost no one notices them anymore. That's true of two moves I observe teachers using for reading instruction in almost every elementary classroom I visit. Both moves—interrupting students to correct their mistakes during oral reading, and asking students low-level questions after they've finished reading—are widespread, despite the fact that no good evidence has ever supported them as effective. At best, both of these moves are unproductive; at worst, they undermine our children's literacy development.
Move 1. Overusing and Misusing Oral Reading
I've been conducting observational studies of classrooms for four decades, and today I observe more oral reading than ever. The sheer volume of oral reading is disturbing, as is the practice of using oral-reading speed and accuracy to make judgments about reading development. In a classic study, Mosenthal (1977) demonstrated that oral reading and silent reading are different processes; a student's skill in oral reading says little about his or her silent reading proficiency, and vice versa.
According to an evaluation of the federal Reading First program (Gamse, Jacob, Horst, Boulay, & Unlu, 2009), our current fascination with oral reading speed has resulted in students who can read aloud faster and more accurately but whose silent reading comprehension has not improved. Given that independent reading with good comprehension is the ultimate goal of literacy instruction, it's puzzling that oral reading activity is so prevalent.
Creating Two Types of Readers
If teachers must continue to use so much oral reading, they should at least reduce its harm by suppressing their tendency to interrupt readers to correct every mistake. The effects of this widespread practice are especially pernicious for struggling readers.
Over 30 years ago, I conducted two observational studies in elementary classrooms, which not only found that oral reading was prevalent, but also that it was used differently with good and with poor readers (Allington, 1980, 1984). One difference was the amount of oral-reading practice that students experienced. Good readers were more likely to read silently during their reading lessons than were struggling readers. Because most people can read much faster silently than they can read aloud, the result was that struggling readers read fewer than half as many words daily as good readers did. This deficit in sheer reading volume is exactly the opposite of what lagging readers need (Torgeson & Hudson, 2006).
Even more troubling than the simple loss of reading practice, though, was the tendency for classroom teachers to interrupt struggling readers both more often and differently than they interrupted good readers. Teachers typically interrupted struggling readers immediately, even before the student had pronounced the whole word that was causing difficulty. In contrast, teachers waited longer before interrupting good readers, usually until the end of the sentence or even the end of the page.
These differences in the timing of interruptions may explain another observation: Teachers tended to correct struggling readers by focusing on surface-level features while encouraging good readers to self-monitor. Consider what happened when a good reader made an error in reading the sentence John went to the store.
GOOD READER: John went to the stone.TEACHER (after the sentence is completed): Does that make sense to you?
The student then reread the sentence, correcting his mistake.
Now consider what happened when a struggling reader misread the same sentence.
STRUGGLING READER: John want—TEACHER (interrupting and pointing at the word went): Look at the vowel in that word.
This interruption led to a bit of unsuccessful word work by the student, followed by the teacher pronouncing the word for him. The student then continued to read.
STRUGGLING READER: … to the story.TEACHER: That e is a silent e. Try it again.
How can we be surprised when these different instructional moves create two different types of readers? Unfortunately, my current observations have found that reading instruction is continuing to separate students into two groups—good readers who self-regulate, and struggling readers who stop after almost every word and look up at their teacher for a cue (Allington, 2012). These differences are not inherent in the struggling readers; rather, they're caused by variations in where teachers direct the students' attention. Good readers learn to pay attention to making sense; struggling readers learn to focus on letters and sounds while paying almost no attention to making sense of what they read.
Refining Oral-Reading Practice
To avoid the harm inherent in the overuse and misuse of oral-reading practice, consider the following recommendations:
- Use oral reading selectively. By the middle of 1st grade, most reading should be done silently.
- If you elect to have students read a text aloud, consciously bite your tongue as they read. Wait until the student has completed at least a full sentence before you interrupt, and then interrupt with a comment that encourages the student to self-regulate.
- Ensure that other students who might be following along or listening to the student read aloud also do not interrupt the reader.
- If you're concerned that you cannot monitor the accuracy of students' reading when they read silently, remember that all you really need to do is ask them to retell what they've read. Misreadings become obvious during retellings.
Move 2. Asking Low-Level Questions
The second misguided but common instructional move that I observe in classrooms is asking an interminable number of low-level, literal questions after (or during) reading. I know that the teacher manuals that accompany commercial reading series are filled with such questions. I'm unsure why, when not a single study demonstrates that this practice actually leads to improved reading comprehension.
Too many of the reading lessons I observe focus on these trivial questions while ignoring how well kids actually understand the text they just read. Sadly, except in a few exemplary classrooms, I almost never witness true literate conversations—the kind that people outside classrooms engage in to make meaning of a text they care about, whether a newspaper article, a memo from the school superintendent, a novel, or a biblical passage.
The Need for Literate Conversations
Imagine that you're sitting in a coffee shop one morning reading the local newspaper when a friend walks in and asks, "Have you read the story about the tornado in Johnsonville?" You respond, "Yes, I just finished it." If your friend were then to subject you to the sort of low-level questions found in most reading series ("What was the fire chief's name?" "What color was the car that was destroyed?") you would probably look at her somewhat grumpily and wonder what was wrong with her. Instead, your friend would be more likely to ask something along the lines of, "That tornado was terrible, wasn't it?" You might respond, "Yes, it was a miracle that nobody was killed!" Your friend might respond with a comment about the article's assessment of Johnsonville's emergency alert system. And thus the literate conversation would begin.
The same sort of literate conversation occurs when someone has read the novel you are currently reading. Two literate adults do not quiz each other on low-level, factual details in the texts they've both read. Instead, they often begin with something like, "How do you like that book?" The literate conversation then follows.
It's unfortunate that our classrooms so often replace literate conversations with interrogations about trivial details. Unfortunate, because we have good evidence that engaging students in literate conversations with their peers is a powerful instructional strategy for fostering both short- and long-term reading comprehension (see Fall, Webb, & Chudowsky, 2000; Malloy & Gambrell, 2011; Nystrand, 2006). Classroom discussions do not need to take up vast amounts of instructional time; research has demonstrated that even brief opportunities for discussion can improve students' understanding of texts and their performance on traditional assessments of reading comprehension (Applebee, Langer, Nystrand, & Gamoran, 2003).
In a study of high-poverty schools, Taylor and colleagues (Taylor, Pearson, Clark, & Walpole, 2000; Taylor, Pearson, Peterson, & Rodriguez, 2003) found that more effective teachers asked five times as many higher-order questions and offered twice as many opportunities for discussion as less effective teachers did. The more effective teachers were also more likely to ask students to respond in writing to higher-order questions. Writing after reading, holding classroom conversations about texts that students have read, and responding to higher-order questions are all linked to higher student achievement. But none of these three instructional moves are routinely observed during classroom reading lessons.
Why Do We Stick with the Trivial?
Given the evidence that low-level interrogation routines are ineffective, why do they continue to be such a common instructional move? One reason may be the current practice of labeling the ability to answer multiple-choice questions on standardized achievement tests as "reading comprehension."
A second factor may be the widespread use of commercial core reading programs that provide almost no suggestions for discussion. Twenty years ago, a colleague and I noted that 98 percent of the questions offered in a commercial reading series were low-level, literal questions (Allington & Weber, 1993). More recent research shows that this proportion seems to be holding true in core reading programs (Dewitz, Jones, & Leahy, 2009).
Third, there's evidence that most teachers are ill-prepared to initiate and manage classroom discussions. Kucan, Hapgood, and Palincsar (2011) found that relatively few elementary teachers were skilled in developing high-quality classroom discussions. Only 15 percent of the teachers they observed could specify the difficulties that students might have with the texts they were given. Most of the teachers did not offer effective support; instead of leading discussions flexibly, they relied on probing for general information and directing students to reread.
Improving Classroom Discussions
This research suggests that teachers must begin to develop their expertise in initiating and managing classroom discussions. Because most students have had little experience with discussion, teachers will likely need to develop students' ability to engage one another as conversational partners.
One instructional move that you can use to do this is turn, pair, and share—having students turn to a student sitting nearby and talk, even briefly, about a text they have just read or listened to. You might initiate turn, pair, and share by providing a specific structure—for example, requiring that one student talk for the first minute of the activity, followed by a minute for the other student, and ending with a minute in which both students are free to take turns talking to each other.
It may also be useful to model how such conversations might proceed and to help students learn appropriate ways to disagree or challenge a response (for example, by saying "I disagree, and here's why"). For teachers worried about the volume of the noise created when multiple pairs are discussing the text, remember that you can model "whisper talk" as an alternative to full, and often loud, conversation.
A specific turn, pair, and share prompt might be to ask students to discuss whether a character in the story reminds them of anyone. Alternatively, you could ask students to discuss their responses to a higher-order question about the text that they have read. For example, when students are reading The One and Only Ivan by Katherine Applegate (HarperCollins, 2012), you might ask, "Do you think that animals really remember things that happened long ago the way Ivan recalled what had happened to his mother and father?" After a few minutes, you can ask one or more pairs to share how their discussion concluded.
Turn, pair, and share enables students to talk through their understandings of what they have been reading. As students develop greater capacity to engage in peer-to-peer discussion, you can ask pairs to jointly write about what they have been discussing. As always, providing a model of what this writing might look like will ease students into this more complex task.
Don't be surprised if many students appear confused or incompetent when you first integrate paired discussions into instruction. Be patient; nothing worthwhile is easy to accomplish. Start with brief turn, pair, and share sessions. Over time, as students become more competent, you can extend sessions and broaden them so the groupings are no longer restricted to pairs but include three to five conversation partners.
Of course, strategies like turn, pair, and share—which enable every student to participate—take more time than teacher-managed discussions in which only a few students are usually involved. But engaging students in literate conversations about what they've been reading must become a common instructional move. You can find time for such discussions by restricting the number of low-level literal questions you ask.
Time to Reconsider
In the end, students are more likely to learn what was taught than to learn what was never taught. Because of schools' overemphasis on oral reading, our students have demonstrated improved oral-reading rates and accuracy but have failed to demonstrate self-regulation or better reading comprehension. Because of schools' failure to make literate conversations a staple of reading instruction, our students daily demonstrate their ability to respond to low-level questions while failing to demonstrate higher-order understanding of what they read. To make literacy instruction more effective, we need to reconsider and fine-tune these common instructional moves.
