Decoding or Fluency Instruction in Middle School?

  • 12 March, 2022

Teacher question

At our middle school, two-thirds of our students are not proficient readers (55% of them read below the 30th percentile). Our staff decided that to help address this problem all our students would engage in repeated reading in each core content class once a week. This would be done with grade-level text. Teachers read the passage, students choral read with teacher, and then students read with a partner. That's it. We have been told that this way of repeated reading will help our students improve their phonics and fluency. I'm not in full agreement with this, but we have been told it is a required whole school intervention. Could you shed some light on this for me and the others who think this is not the best bang for our buck? What about explicit phonics instruction?

Shanahan response:

This is really two questions: one about the best way to meet secondary students’ reading needs and one specifically about the value of teaching fluency. I’ll try to answer both.

As for your middle school’s unfortunate reading levels, fluency instruction may be part of the solution, but it will probably be insufficient, especially if implemented in the way that you describe.

That 55% of your students are below the 30th percentile on some reading test, tells me that you have a serious problem, and such problems deserve a serious response from serious people.

But those test data don’t reveal what the problem is.

There are many reasons students in grades 6-8 may not read well. Decoding problems, perhaps. Or difficulties with fluency. Poor vocabulary and weak morphological skills? Maybe the students lack reading stamina, knowledge of written language, or adequate reasoning ability. Maybe, they can read, but don’t want to. And, of course, the problem could be some mix of all those possibilities.

Years ago, I would have said, “avoid phonics instruction.” Studies then had concluded that phonics at those grade levels didn’t pay off.

Then, I was on the National Reading Panel, and we found studies that showed phonics instruction could improve the decoding abilities of older students but without positive impacts on reading comprehension or spelling. I took that to mean that it was reasonable to address the phonics needs of older students in pull out interventions, but that phonics alone would be pointless (these kids would need vocabulary and comprehension – a conclusion consistent, I thought, with the work of Charles Perfetti who claimed that there was more to reading words than decoding). Accordingly, I invited Don Deshler to provide advice in this space about high school phonics instruction, and Marilyn Adams chimed in on the success of her program for doing just that (you can look it up).

More recent studies have reshaped my thinking again. One study has been especially influential in that regard (Wang, Sabatini, O’Reilly, & Weeks, 2019). They administered reading tests to tens of thousands of students in Grades 5-10. Not a nationally representative sample and not a widely used battery of tests either (Reading Inventory and Scholastic Evaluation – RISE). Nevertheless, the findings are provocative.

This study reported that a significant percentage of the middle and high school students had very low levels of decoding ability – 30% of the students in their sample, though it should be remembered that this was an especially low reading population.

The cut point or threshold they used to identify this 30% was interesting. They found that especially low decoding ability didn’t correlate with reading comprehension. When students got above a certain point in decoding, there was a correlation.

They divided their sample of students at that inflection point and that’s where things got interesting.

The students who scored below that point made essentially no learning progress in reading across the school year, no matter the students’ grade level.

The percentage of students in need of explicit decoding instruction in your school may be lower than in that study. Nevertheless, I’d worry a lot about those kids who perform below that decoding threshold. I would be surprised if they got much from the fluency training that you describe.  

The original theory of oral reading fluency instruction (which I attribute to Carol Chomsky) was that some students who had adequate knowledge of the orthographic-phonemic system couldn’t implement it during reading. Fluency instruction was for those children who tested high in phonics but low in reading.

The Wang et al. study suggests to me that Chomsky may have been on to something. Students below the decoding threshold would gain little or nothing from fluency work, while those above it could benefit. The National Reading Panel found that fluency practice had a positive impact on reading comprehension, but even a bigger payoff for word reading, including decoding. That is, the kind of fluency practice you describe can improve decoding ability (perhaps because it helps students to develop automaticity with those phonics skills).

The pattern identified in the Wang study seems to fit with a lot of other facts. For instance, it might explain why it has been so difficult to obtain positive results from secondary school reading interventions. Those studies usually lump all kinds of low readers together. The presence of the below threshold readers could exert a misleading drag on the effects of vocabulary, language, or strategy interventions – since the low decoders wouldn’t be able to benefit yet from such teaching.

Also, even if phonics interventions for older students targeted those below-threshold kids specifically, the results would likely be disappointing. Remember, students didn’t benefit from other instruction until they exceeded the threshold. That would mean that reading gains would not be likely until sometime after the intervention was completed, long after typical post-testing in research studies (or intervention programs).

In any event, those results suggest that some of your students will probably require explicit decoding instruction. Further testing would be needed to determine who they were. Fluency practice can have positive impacts on decoding but perhaps mainly for students who are at a high enough level of decoding ability.

Other students might need greater attention to vocabulary, morphology, written language analysis, comprehension strategy work, and/or world/domain knowledge development. Fluency practice won’t help with any of those deficiencies either.

Fluency instruction has been found to improve reading achievement, so providing some fluency support should have some value (though what you describe isn’t encouraging).

Modeling can be helpful (a teacher reading the text prior to the students’ attempt to read it) but reading a whole passage doesn’t provide much memory support. I’d suggest limiting that modeling, making it briefer and more targeted. If the students have trouble with a sentence, read that to them, and then have them give it a try.

Paired reading can be helpful, too, if the other students provide feedback. Teachers need to involve themselves in this process, since not all students are well-equipped for it. I recommend pairing students strategically to make sure that those struggling get the needed support or that no student always gets stuck with the partners who won’t or can’t help.

I would encourage all the teachers to use these practices in class but give them control over when they do this. Perhaps everyone should sign onto a certain number of minutes of fluency each week (and make it enough minutes that the students will get several such experiences weekly).

Finally, motivation is important. Explaining the point of fluency work to the students matters. Tim Rasinski strongly encourages activities like Radio Reading and Readers’ Theatre as engaging versions of fluency. Such motivation can be overdone in my opinion (because it can take too much time), but goosing things up occasionally makes sense.

The simple answer is that fluency practice can be beneficial for your students. It will help those who have fluency problems. But what’s going to be done with the kids with other reading needs?


Chomsky, C. (1976). After decoding: What? Language Arts, 53 (3), 288-296, 314.

Wang, Z., Sabatini, J., O’Reilly, T., & Weeks, J. (2019). Decoding and reading comprehension: A test of the decoding threshold hypothesis. Journal of Educational Psychology, 111(3), 387-401.


See what others have to say about this topic.

Mar 12, 2022 06:08 PM

I'm a little troubled by this answer because the reading together and engaging in text no matter what would seem more valuable than a lot of what students do now, like watch a video or draw. I would posit many of these kids got to this place because they were doing much lower value activities in class and I don't see the harm in the shared reading??

Jacqueline Leigh
Mar 12, 2022 06:10 PM

I don't see it mentioned here that some of the students in any of the situations you were describing would be at some stage of learning English as an additional language.

Mar 12, 2022 06:17 PM

This makes so much sense to me. How might we assess the other areas (vocabulary, morphology, written language analysis, comprehension strategy work, and/or world/domain knowledge development) to determine their exact needs? Intervention often just target all of those areas and I sometimes wonder about how efficient we really are. The assessments for phonics and phonemic awareness are so concrete and get a lot of coverage. How do we assess these other areas?

Mark Anderson
Mar 12, 2022 06:38 PM

This is excellent! One thing I paused on, though: "Other students might need greater attention to vocabulary, morphology, written language analysis, comprehension strategy work, and/or world/domain knowledge development. Fluency practice won’t help with any of those deficiencies either."

It seems like methods that combine multiple texts aligned by topic show promise for tying in word/domain knowledge developing alongside the fluency practice -- methods like "Wide FORI" and "Varied Practice." By reading multiple texts that build conceptual knowledge and vocabulary, and tucking in fluency practices like repeated echo and choral readings after a 1st modeled read, it seems like fluency practice can do some good work to build knowledge and vocabulary -- and even potential morphology and comprehension strategy work, if close reading for language and meaning is conducted alongside of the fluency practice. More research along these lines needed, of course.

Timothy Shanahan
Mar 12, 2022 06:45 PM


That's a real problem with a lot of current assessments... they do a good job with foundational skills (which ain't nothing), but there aren't quick and easy measures of depth of understanding and the like. Though it looks like a formidable battery of tests, the tests mentioned in this blog sound exciting to me -- as they measure many features of reading at these grade levels. I don't know how practical these tests are in terms of time and cost, but perhaps they'll help fill this bill.


Timothy Shanahan
Mar 12, 2022 06:49 PM


You are correct, I did not mention anything about special populations, including English Learners. However, the research is pretty clear that ELs need to learn the same reading skills as native speakers -- though the normative performance on various skills will usually differ. For example, ELs are no more likely than their first language peers to need a phonics intervention, but they are much more likely to need a vocabulary and oral English one. Likewise, the payoffs from instruction in decoding, fluency, and reading comprehension strategies has a lower payoff than when provided for the native English speakers -- but the payoff from language interventions is higher. It is unknown at this time whether the threshold theory described here applies to second language learners, but I can't think of any reason why it would not.


Timothy Shanahan
Mar 12, 2022 06:53 PM

The point of the blog was to focus on instruction that would be beneficial in improving middle schoolers reading ability. The possibility that there is a decoding threshold that must be accomplished before such lessons will stimulate learning is important -- especially when that is the only effort being made to address these students' reading needs. Knowing that you prefer what for those particular kids would be a waste of time over other activities that wouldn't meet their needs is not heartening. I'm a big proponent of explicit fluency instruction -- but I'd target it on the kids likely to benefit and I'd try to do it in the most effective ways possible.


Mar 12, 2022 08:22 PM

It seems like in our state we are heading in the right direction with this. If you screen your kids at middle school who are struggling with reading, usually a very low percentage (at least at my school) have problems with phonics, BUT if they do, they are obviously in need of intensive intervention. Our new guidelines are that these students must get an FIE. That means as instructors, we get back valuable data about all of the reading areas in which the student struggles, including vocabulary, spelling, comprehension, etc. From there, we can target the intervention appropriately. Without the data from quality testing, it would be hard to determine what those low level readers are really struggling with.

Timothy Shanahan
Mar 12, 2022 08:27 PM


That sounds great to me. However, the study discussed here scares me a bit because it could mean that some kids that you and I would currently see as having accomplished sufficient phonics may be below that threshold. Too early to know, but I guess the trick for now is to continue to monitor those kids who came the closest to being assigned to phonics (but who were not) to make sure they are making progress with what you are providing them.


Dr. Bill Conrad
Mar 12, 2022 09:17 PM

Excellent linkage between the research with the practical work teachers do day to day! Well done!

Tim Rasinski
Mar 12, 2022 09:54 PM

HI Tim. Excellent blog as always. I wonder why the answer to the question you pose in your title could not be "both." Instruction can be designed so that word study/decoding instruction can be woven into fluency. In the Fluency Development Lesson (Rasinski, Padak, Linek & Sturtevant, 1992) that the National Reading Panel included in its report word study was an integral component that followed the assisted and repeated reading portions of the lesson.

Stacy D.
Mar 13, 2022 01:56 AM

Before sharing my thoughts, I'd like to thank you, Dr. Shanahan, for taking time to answer questions, open discussions, and lead us to improve our understanding and use of research and evidence based practices. With that being said, I was very intrigued by the idea of the "decoding threshold," and how I could leverage that to best meet the needs of my students. As an elementary reading interventionist, the data I collect formatively through exit tickets, reading records, and classroom data, drives everything I do to meet their needs. In grades K-3, the intervention focus on decoding has had a direct positive impact on their overall reading skills, resulting in a constant adjustment to the groups I serve. But, in grades 4-5, the impact is not as obvious, and I often feel like I'm "not doing my job" to put it simply.

I followed your citations to the Decoding Threshold Hypothesis in ERIC and began reading. Everything made sense as to why we are not seeing a direct impact on many of the overall reading scores when measured against their peers, even when we see positive results on formative decoding assessments. So now I have new questions. What IS the decoding threshold? A simple answer appears to be that it will differ, as some students have "graduated" from my services. Is there a "magical list of skills" that might help us approach this threshold faster? Is the explicit instruction that I'm providing missing a key component for the older students, and what is it?

Thank you for all you do to support teachers and practicioners in education, and thank you for your time!

Timothy Shanahan
Mar 13, 2022 04:38 AM


I share your excitement. But remember at this stage we don't have definitive proof that there is a particular threshold. If you are using the test from the study, you could easily set the same threshold that they found for each of 6 grade levels, but i don't know at this point how that could be done with any other test without further research. Exciting idea, makes sense, we need more data to be sure.


Mar 13, 2022 11:49 AM

I taught middle and high school reading for years (I’m currently at the elementary level). I was discouraged with all the phonics instruction. Kids do not like feeling like a 5 year old. When I gave a 1 minute fluency assessment I found kids read accurately but rate and expression were the major issue and I had to prioritize instructional focus. I discovered Rasinski’s work and it helped my teaching so much. Partner reading and rereading improved skills and understanding tremendously. My students were pretty much all low but made major improvements because of engagement with his techniques. Partner reading always includes a W question for the listener at a stopping point. I do this at the elementary level as well with grades 3-4 (with some phonics too). What is amazing for me to see is that something kicks in with word recognition by 4th grade. I’ve serviced the kids in grade 1, 2, 3 that have received AIS support and by 4th they recognize words with much better accuracy and rates. By 4th instruction is mostly fluency and comprehension.

Ed Jones
Mar 13, 2022 01:29 PM

I’d point to this conversation with Ms. Jasmine then of Minneapolis schools. (Hardened BL curricula.)
Her good reading teens were thrilled to find that, wow, there is a magic process for decoding words.

Age-appropriate content is key, of course. But, presented in a way that shows the deficit is in no way the teens fault, but rather the district leadership, they’ll embrace and dig in, as they might with practicing foul shots or musical exercises.

Laura Johnson
Mar 13, 2022 04:17 PM

I am so glad I read this blog today! This information has been so reaffirming for what we are starting at my high school.

I am seeing the same struggles with students. In my second year as a literacy specialist at the high school level, I have been able to help to revamp the RTI process for literacy. We now use the NWEA Map Assessment as a universal screener for our Freshmen and Sophomores. I review the data, and then decide where to begin administering diagnostic assessments. This year I have been able to assess students who fell below the 35%. I use a grade level ORF and the San Diego Quick Assessment of Reading. These results then lead me to decide if I should go further with an advanced and beginning decoding survey. The data from these diagnostic assessments drives how I form my intervention groups. At this time my Freshmen and Sophomore groups are getting advanced phonics instruction and morphology in an intervention.

Intervention occurs in a few ways, in two Tier II support classes for Freshmen and Sophomores and during lunch periods. Attendance for lunch groups has been a challenge, but those who are attending are seeing gains.

This semester I had the opportunity to present the Map reading data to a Freshmen team of teachers (across all content areas), and we explored trends in the reading data. From this in-house PD, we have begun to create a Tier I “Freshmen Focus” on vocabulary instruction. So far, it looks like teachers identifying “must know” Tier II and Tier III words for their current units of study. Some PLCs have gotten as far as including explicit instruction and multiple encounters with vocabulary within their subject matter. I feel this is a pretty good start—especially with the challenges of pandemic teaching and learning.

Thanks for writing this blog and sharing your literacy knowledge with all of us! If you come across other high school models of literacy intervention, please share :)

Jo Anne Gross
Mar 13, 2022 10:09 PM

Thank You for opening the rat’s nest!
All the crappy instruction deployed on students by the flawed instruction perpetuated by the “balanced” epoch is sitting in those kids.
They’re fuzzy, they’re blurry.
They’re blocked from school comfort and the feeling of competence!
Seen too many to count.

Mark Pennington
Mar 14, 2022 12:45 AM

While I commend the school for encouraging a shared responsibility for literacy, I agree with you, Tim, that a one-size-fits all approach to fluency practice without diagnosing and treating the underlying causes of poor fluency among many of the students is probably not worthwhile.

However, with a few tweaks this notion of schoolwide fluency practice could be quite effective for most students. If for example, students completed simple fluency assessments (perhaps in each content area), teachers could use the data to determine appropriate reading rates at students' challenge levels. In my experience, 20% faster than the assessed fluency rate seems to work as a good challenge rate.

Teachers could print a weekly content passage of about 400 words with word counts in the margin and record the passage at, say, three levels of speed. Note, using the free app, Audacity, changes the speed, but not the pitch, so teachers would have to record just one reading.

The teacher could assign the appropriate challenge levels to students, and students could complete repeated readings along with the teacher's recordings. Teachers may wish to record unpracticed and practiced timings to motivate students and track progress.

Miss Emma
Mar 14, 2022 11:59 AM

As you know, of course, word processes that are inefficient hinder text understanding by reducing the cognitive resources available for comprehension. Slow, non-automatic word recognition processes in alphabetic orthographies typically occur when children are faced with trying to ‘sound out’ most of the words, or by laboriously using sentence context cues to guess the unknown word. They haven’t moved into the ‘sight word' reading phase (Ehri 2014)

The problem with ‘teaching phonics’ as a solution is that ‘phonics’ is viewed so narrowly, and as if only learned with explicit instruction. They generally only cover 100 or so phoneme to grapheme correspondences, and even ignore the Schwa to grapheme mapping eg caterpillar would be ‘sounded out’ using the correspondences taught, and they ‘translate’ this if they know the word. They don’t then explore the actual speech sounds to grapheme mapping Ie the phonics program ignores the issue of English not having a transparent orthography. They teach some common correspondence, usually ‘sight words’ as whole words (ignoring the mapping) and think if the kids then repeat reading they’ll somehow decode with fluency and comprehension.

As you said, no point working on ‘fluency’ if decoding skills are weak. Like a baker working on piping skills to decorate the cake when every cake they bake sinks.

I’ve developed techniques that our teachers use in upper primary and with high school kids still stuck on that single word decoding/ reading phase. They work hard to understand what the kids need, and so the learning experiences centre on the kids. The focus is on what they’re learning - more than what the teacher should ‘plan’ for a group or class. It’s all differentiated.
And that’s what’s hard - many teachers following programs aren’t really sure about even the phoneme to grapheme mapping that underlies orthographic mapping. It might be why they stick to basic phonics programs. They stink to activities and texts that are ‘decodable’ using a limit set of correspondences.
A lot of kids will learn with that ‘kick start’ but many won’t. That group being discussed here, for example.

Those basic ‘kick start’ programs also tend to ignore accents, and often even a failure to address phonemic awareness deficits and weak oral language skills.
We need to give teachers more knowledge rather than keep churning out programs. We need teachers to understand the cognitive processes so that they look at each individual to find out why they’re stuck someone along the learning to read (and spell) continuum. The ‘kick start’ has to also offer opportunities for implicit learning and awareness of phoneme to grapheme mapping (speech to print can help) and allow kids to learn what they need, when they need it. That takes a skilful teacher. We know what kids are likely to need - at different points - but at least 65% of kids need a teacher who offers that, as it won’t happen for them without that support.

Munro Richardson
Mar 14, 2022 12:19 PM

I run a countywide early literacy initiative in Charlotte, North Carolina. We've done a lot of work focused on reading fluency since 2018, sometimes referred to as the "forgotten literacy skill." We have successfully used a wonderful evidence-based fluency intervention called HELPS. It has 5 small RCTs and is included in the WWC practice guide for K-3 foundational skills.
HELPS is available for free at . HELPS elegantly integrates 7 or 8 research-based practices into one intervention (modeled reading, repeated reading, etc.) I myself was a HELPS tutor during the first year of roll out in the 2018-2019 school year and saw how well it worked with the two ELL students I tutored.

As it relates to this conversation, HELPS sets a minimum floor for words correct per minute (WCPM) by grade level and time of year (BOY, MOY and EOY). Students who do not meet the minimum level of reading fluency are not likely to benefit from this program -- they need help with phonics (and possibly phonemic awareness). Although the focus of HELPS is elementary grades (starting with mid-first grade), I know of groups that have been able to use the intervention with middle school students.

My point is not to promote an intervention -- although HELPS is a good one -- but rather to reinforce the point about the distinction between those students who need phonics-based tutoring versus those that may benefit from a fluency-based intervention.

Rebecca Carlson
Mar 17, 2022 01:39 PM

My concern is that the past 2 years of disrupted education and the hesitancy of many of our students to read we will see the gaps widen even greater in the future.

Students are bombarded with other options due to tech advancements. Some schools are shifting attention and funds towards tech and away from books and reading as a core mode of teaching. There shouldn't be an abandonment of libraries to bring in tech but it's the trend for many districts at the expense of the literacy of our students.

Claude Goldenberg
Apr 09, 2022 01:07 PM


You might have seen this opinion piece I did in Ed Week a few months ago.

I cited Joe Torgesen’s under-cited 2004 American Educator paper where he argued persuasively that we have the tools and wherewithal (but apparently not the will) to help as many as 95% of all kids get at least to what is now the 30th percentile ("low average” in his appraisal) in basic reading skills by end of grade 2.

However, he also said that we lack such knowledge and tools from grade 3 and beyond, since hitting at least low average in reading skills after grade 2 requires a whole lot more than the foundational skills central to beginning and very early reading. He challenges the reading community, especially researchers, to do the work to enable us to hit that mark from grade 3 and up. Nothing I’ve seen nor anyone I’ve asked indicates we’re close to meeting the “Torgesen challenge.”

What do you think? Are we any closer than we were nearly 20 years ago to meeting this challenge?

Claude Goldenberg
Stanford University

Timothy E Shanahan
Apr 09, 2022 01:55 PM


Thanks for this. I believe that things have not moved as fast as Joe would have hoped and, yet, there have been real advances on both the foundational skills side (such as the study emphasized here) and on the language/comprehension side. There have been a number of studies exploring how to improve reading through an instructional emphasis on the linguistic aspects of reading (e.g., vocabulary, syntax, cohesion, discourse structure) and on knowledge. Likewise, there has been a growing interest in teaching students to read with texts that they cannot already read well -- a real difference in what we have been doing and real difference with how beginning reading seems to work.

Despite increases in knowledge, I see no serious change in how reading is taught at those upper grade levels at this point. Too many people have bought in on the idea that if kids read well early then the job is done. Winning the first half of a game shouldn't be the goal. Scientists certainly could make faster progress at these levels, but their rate of success only matters if schools adopt practices based on that increase in knowledge.

thanks for this.


Apr 12, 2022 05:33 PM

Your idea seems like a hopeful suggestion because of the fluency/comprehension connection. Repeated readings are meant to increase a student's reading rate which can improve comprehension. Your kids will most likely improve their reading rate, but the problem is, they are memorizing or "parroting," and not actually reading.

A 30% decoding level shows your kiddos aren't fluent ---- at decoding.

Data identified the need as decoding. We rationalize it, offer a "doable" intervention, and ignore the identified problem. It's funny. It's like painting a sinking boat.

What Are your thoughts?

Leave me a comment and I would like to have a discussion with you!

Comment *

Decoding or Fluency Instruction in Middle School?


One of the world’s premier literacy educators.

He studies reading and writing across all ages and abilities. Feel free to contact him.