How to Teach Fluency So That It Takes

Blast from the Past: This entry was first posted on September 17, 2017 and was re-posted on March 2, 2019. This entry isn’t from very long ago, but the guidance it provides appears to be timely given some of the recent questions that I have received. Test preparation season seems to be in full swing, so many of my correspondents are questioning both the value and the approaches taken to oral reading fluency. Research is clear—many of the boys and girls who don’t do well on their state tests are disfluent and that undermines their ability to comprehend well. Oral reading fluency is not tested directly on those tests, but it is indexed. This enty explains how to teach fluency effectively--which should enhance students' silent reading comprehension.

Teacher question:

I have a question regarding my school's reading program. My question today is about the reading portion of our literacy block and most specifically the partner reading and independent reading.

I'm finding that my homogenous group of fourth-grade students aren’t fluent readers. The routine expectation is that partners take turns reading a paragraph at a time. The partner who is following along and not reading aloud is to provide a brief summary of what was read by the partner before reading the next paragraph. I love this, except that my students aren't fluent readers, so I feel that first the comprehension is low because of non-fluent reading, and second the time is a bit wasted because of the lack of fluency and therefore comprehension. After students do their partner reading, they read the next couple pages independently. Again they aren't fluent, so it's taking quite awhile. I feel that comprehension is low.  

Shanahan response:

Fluency instruction can be valuable with fourth-graders (and with lots of other kids in grades 1-12)—it can help them to decode better, read more fluently, and improve reading comprehension.

What you describe is not likely to have much impact on kids’ learning. Fluency instruction requires, well, instruction.

Text selection. Good reading instruction requires appropriate texts. On this, it sounds like you’re doing fine—that often isn't the case. Fluency practice is best carried out with texts that students will struggle with, and with fourth-graders this is most likely to mean that the kids won’t recognize all the words.

I would shoot for texts in which kids would make about 10 mistakes per hundred words. Given that your boys and girls are struggling to read the texts well enough to understand them, it sounds like you are using texts are hard enough to be effective for this teaching.

But while you want to use texts that difficult for these lessons… you don’t want to end there.

If kids aren’t reading the text fluently by the end of the lesson, then the lesson itself is ineffective. (Unlike you, many teachers try to teach fluency with texts that are too easy. That doesn’t help the kids learn anything because they are already sufficiently fluent when they start.)

Purpose setting. Fluency lessons are no different from any others; students need to understand the purposes of the lesson.

I usually start out with a discussion of what oral reading should sound like. You might consider making a poster or bulletin board based on the kids’ insights about oral reading. Perhaps read a text to them to show what reading should sound like, and then do some less than terrific readings (e.g., making mistakes, reading too slow or too fast, pausing badly, reading choppily).

Tell kids upfront that you are starting them with texts that they will probably not read well the first time through. Explain that you want them to work on that text until they can read it well.

Before each day’s fluency work, remind the kids what they are trying to accomplish and what reading should sound like.

Modeling. There are some studies of fluency instruction in which teachers read the texts to the students before the students do their own oral reading. That can be helpful because it tips kids off to some of the unknown words and gives them some clues about the content. (Sometimes audio recording provides the modeling).

However, there are also many studies in which there is no such modeling.

I tend to put myself in the “don’t model” group. Experience tells me that it only helps if you read a very short portion of the text, like a sentence, and then immediately have kids try to read the same sentence. Verbatim memory in these circumstances doesn’t last longer than that, so modeling tends not to be a big help.

If a student tries to read something and makes a real mess of it (e.g., lots of mistakes, poor phrasing, etc.), then I might read that sentence aloud and have the child try it again. In other words, I’d provide modeling only when I think the students are struggling so much with a sentence that they are not likely to improve on a second (or third) reading without that extra boost.  

Partners. Research suggests that lots of people can be effective reading partners. Teachers, parents, volunteers, cross-age tutors, computers, and classmates have all been tried in one study or another and they have all been effective. In your case, you have kids partnered up, but you told me nothing about how these pairings were made and what training was provided to them to enable them to be effective.

Studies (and personal experience) tell me that not all kids are great partners. Some are careful and supportive and others could care less. If you get Mariel as a partner, there is a good chance that your reading will improve; if you’re paired with Bobby—not so much.

In some studies, children are paired on the basis of ability, and that can be successful, at least generally. In such pairings, the bottom students are placed with the top ones. This makes sense: The better readers can certainly help with errors, but that means they get no help and just because someone is a good reader, that doesn’t make them patient or helpful.

I believe in sharing the pain, sharing the gain.

Match kids with a different student partner every day. That way, if getting to work with Mariel is a great opportunity, everyone will enjoy that opportunity every few weeks. If Bobby is a dud, everyone suffers that occasionally, too.

Feedback. Partners should be trained to give beneficial feedback. In all of the studies that have found fluency instruction to have a positive impact on reading, the readers received feedback.

I’m a fan of the “Pause, Prompt, Praise” system. If a student errs, the partner waits until the end of the phrase or sentence. The idea of that is to give the reader a chance to fix the mistake. If the reader doesn't take it on, then the partner should stop him/her.

The partner should then try to help by offering a prompt. There are three kinds of prompts: (1) encourage the reader to sound out the word (e.g., look at that more closely; sound it out; try to break it apart, etc.); (2) encourage the reader to use meaning to figure out the word (e.g., does that make sense? what should that say?). Or, (3) you can tell the reader the word.

Never give more than two cues. If the reader doesn’t get the word right, then tell him/her the word.

If neither partner knows the word, they should write it down and the teacher can guide then guide the whole group to decode these words at lesson’s end.

Rereading. Your letter does not mention it, but research suggests that the single most important step in fluency practice is rereading. Your students are reading a text badly, and your lesson seems to go on.

What should be happening is that the students should be asked to reread those portions of the text again. And, they should sometimes do so even a third time.

That’s where the learning is. By reading and rereading a text the students transform it from one they can’t read well to one that they have read well. In some schemes, they read the text again and again until they can do so, and in others, 3 readings seem to be sufficient.

Having kids taking turns reading paragraphs aloud poorly is not effective teaching. Whereas, having kids read and re-read a text gives them an opportunity to figure out unknown words and how to make the text sound meaningful.

Teacher’s role. The teacher plays a very important role in this process. Teachers should supervise these paired readings. Teachers should coach readers through the partners (this is part of partner training). For example, I would eavesdrop on a pair, listen to the reader read and wait to see what the partner does. If the reading was weak I’d ask the partner, “What do you want her to do?” or tell the partner, “I’d tell him to read that again.”

Make your way around the group, making sure that you hear several kids reading each lesson and that you add your feedback and modeling as needed.

Comprehension. In some studies, comprehension is handled in the way you described. In others, comprehension is not dealt with a fluency lesson. It might not be absolutely necessary, but I would encourage you to continue with the summarization practice as described. But have the students delay doing it until the paragraph has been read reasonably fluently.

Amount of Fluency Instruction. Schedule anywhere from 30 minutes to 45 minutes per day for fluency teaching (about 25% of the reading instruction time).

If kids finish the text prior to the end of the lesson time, then have them switch places, reading the paragraphs that their partners read the first time.

If you want kids to learn from fluency instruction, you need to teach fluency. Try that and let us know how it goes.