Showing posts with label Oral Reading Fluency. Show all posts
Showing posts with label Oral Reading Fluency. Show all posts

Thursday, October 27, 2016

Oral Reading Fluency is More than Speed

Letter I received:

I found these troubling quotes in the Report of the National Reading Panel:

"Fluency, the ability to read a text quickly, accurately, and with proper expression..."

"Fluent readers can read text with speed, accuracy, and proper expression..."

My dismay is due to (a) listing rate first in both statements, and (b) using "quickly" and "with speed" rather than "rate" (or "appropriate rate" as in the CCSS fluency standard). I wonder if this wording may have encouraged folks who now embrace the notion that "faster is better" (e.g. "better readers have higher DIBELS scores--wcpm")

In my own work I often refer to Stahl & Kuhn (2002) who stated that "fluent reading sounds like speech"-- smooth, effortless, but not "as fast as you can."

Who’s right?

Shanahan response:

            Well, first off, let me take full responsibility for the wordings that you found troubling. I took the lead in writing that portion of the report, and so I probably wrote it that way. Nevertheless, I doubt that my inapt wording was what triggered the all too prevalent emphasis on speed over everything else in fluency; that I’d pin on misinterpretations of DIBELS.

            I, too, have seen teachers guiding kids to read as fast as they can, trying to inflate DIBELS scores in meaningless ways. What a waste of time.

            But, that said, the importance of speed/quickness/rate in fluency cannot be overstated—though it obviously can be misunderstood.

            The fundamental idea that I was expressing in those quotes was that students must get to the point where they can recognize/decode words with enough facility that they will be able to read the author's words with something like the speed and prosody of language. 

            Old measures of fluency—like informal reading inventories--looked at accuracy alone, which is only adequate with beginning readers. The problem with accuracy measures is that they overrate the plodders who can slowly and laboriously get the words right (as if they were reading a meaningless list of random words). 

            DIBELS was an important advance over that because it included rate and accuracy--which is sufficient in the primary grades, but which overrates the hurried readers who can speed through texts without appropriate expression. Studies are showing that prosody is not particularly discriminating in the earlier grades, but as kids progress it gains in importance (probably because the syntax gets more complex and prosody or expression is an indicator of how well kids are sorting that out—rather than just decoding quickly enough to allow comprehension).

            Fluency instruction and monitoring are very important, and I agree with your complaint that it is often poorly taught and mis-assessed by teachers. I think there are a couple of reasons for that.

            First, I think many teachers don’t have a clear fluency concept—and stating its components—accuracy, rate, and prosody—in their order of development won’t fix that. Fluency is not a distinct skill as much as it is an amalgam of skills. It is part decoding, part comprehension.

            Kids cannot read if they can’t decode and recognize words; translating from print to pronunciation. That’s why we teach things like sight words, phonological awareness, and phonics.

            However, recognizing words in a list is a very different task than reading them horizontally, organized into sentences, with all the distraction that implies. Speed (or rate or quickness) don’t really matter when reading a list of words. But when reading sentences, it is critical that you move it along. Slow word reading indicates that a student is devoting a lot of cognitive resources to figuring out the words, and that means cognitive resources will not be available to thinking about the ideas. That’s why speed of word reading is so important; it is an indicator of how much a reader will be able to focus on a text’s meaning.

            But fluency is not just fast word reading. It includes some aspects of reading comprehension, too. For instance, fluent readers tend to pronounce homographs (heteronyms)—desert, affect, intimate—correctly without needing to slow down or try alternatives. Fluent readers may have no advantage in thinking deeply about the ideas in a text, but they do when it comes to this kind of immediate interpretation while reading.

            Another aspect of comprehension that is part of fluency is the ability to parse sentences so that they sound like sentences. Someone listening to your oral reading should be able to understand the message, because you would have grouped the words appropriately into phrases and clauses. To read in that way, you, again, have to be quickly interpreting the sentences—using punctuation and meaning as you go.  

            Teachers who think that fluency is just reading the right words, or just reading the right words really fast, is missing the point. Stahl and Kuhn are right: fluency has to go, not necessarily fast, but the speed of normal language.

             Second, I think many teachers don’t understand assessment. Reading assessments of all kinds try to estimate student performance based on small samples of behavior. Accordingly, the assessment tasks usually differ from the overall behavior in important ways. With fluency that means measuring some aspects of the concept—speed and accuracy—while not measuring others—prosody.

            Given the imperfect nature of these predictor tasks, it is foolish, and even damaging, to teach the tasks rather than the ability we are trying to estimate. It is like teaching kids to answer multiple-choice questions rather than teaching them to think about the ideas in text.

            As long as teachers try to teach facets of tests rather than reading we're going to see this kind of problem. The following guidance might help.

1.    Tell students to read the text aloud as well as they can—not as fast as they can.
2.    Tell them that they will be expected to answer questions about the text when they finish—so they will read while trying to understand the text.
3.    Pay attention not just to the wcpm (words correct per minute), but to whether the reading sounds like language.


November Powerpoints

-->

Wednesday, February 3, 2016

Are Oral Reading Norms Accurate with Complex Text?

Teacher Question:  
          A question has come up that I don't know how to address and I would love your input.  For years, we have used the Hasbrook/Tindal fluency norms as one of the ways we measure our student's reading progress.  For example, the 4th grade midyear 50th percentile is 112 CWPM.  The fourth grade team has chosen a mid-year running record passage and is finding that many of students have gone down instead of up in their CWPM.  One teacher said that is because the common-core aligned texts are more challenging and that the passage is really the equivalent of what used to be considered a 5th grade passage. She said that the norms were done using text that is easier than what the students are now expected to read. I know that the texts are more complex and challenging and therefore more difficult for the students to read, and that this particular text may not be a good choice to use for an assessment, But it does raise the larger question--are these fluency norms still applicable?

Shanahan response:
         This is a great question, and one that I must admit I hadn’t thought about before you raised it. If average fourth-graders read texts at about 112 words correct per minute by mid-fourth-grade, one would think that their accuracy and/or speed would be affected if they were then asked to read texts that in the past would have been in the fifth-grade curriculum.

         However, while that assumption seems to make sense, it would depend on how those norms were originally established. Were kids asked to read texts characteristic of their grade levels at particular times of the year or was the text agenda wider than that? If the latter, then the complex text changes we are going through would not necessarily matter very much.

         So what’s the answer to your question? I contacted Jan Hasbrouck, the grand lady herself, and put your question to her. Here is her response:

         I guess the most honest answer is "who knows?" I hope that we may actually have an answer to that question by this spring or summer because Jerry Tindal and I are in the process of collecting ORF data to create a new set of norms, which should reflect more current classroom  practice.

         My prediction is that the new ORF norms won't change much from our 2006 norms (or our 1992 norms). My prediction is based on the fact that ORF is, outside of expected measurement error (which Christ & Coolong-Chaffin, 2007 suggest is in the range of 5 wcpm for grades 1 and 2 and 9 wcpm in grades 3-8+), fairly stable. You can see evidence of this on our 2006 norms when looking at the spring 50th %iles for grades 6 (150), grade 7 (150), and grade 8 (151). When you think that these three scores represent approximately 30,000 students reading a variety of grade level passages that pretty darn stable. Other studies of older readers (high school; college) also find that 150 wcpm is a common "average.”

         Of course this stability assumes that the ORF scores were obtained correctly, using the required standardized procedures, which unfortunately is too often not the case. Standardized ORF procedures require that students read aloud for 60 seconds from unpracticed samples of grade level passages, and the performance is scored using the standardized procedures for counting errors. In my experience most educators are doing these required steps correctly. However, I see widespread errors being made in another step in the required ORF protocol: Students must try to do their best reading (NOT their fastest reading)!  In other words, in an ORF assessment the student should be attempting to read the text in a manner that mirrors normal, spoken speech (Stahl & Kuhn, 2002) and with attention to the meaning of the text. 

         What I witness in schools (and hear about from teachers, specialists, and administrators in the field) is that students are being allowed and even encouraged to read as fast as they can during ORF assessments, completely invalidating the assessment. The current (2006) Hasbrouck & Tindal norms were collected before the widespread and misguided push to ever faster reading.  It remains to be seen if students are in fact reading faster. Other data, including NAEP data, suggests that U.S. students are not reading "better."

         And yes, of course the number of words read correctly per minute (wcpm) would be affected if students were asked to read text that is very easy for them or very difficult, but again, ORF is a standardized measure that can serve as an indicator of reading proficiency. 

         Given Jan's response, I assume the norms won’t change much. The reason for this is that they don’t have tight control of the data collection—reading procedures and texts varying across sites (not surprising with data on 250,000 readers). That means that the current norms do not necessarily reflect the reading of a single level of difficulty, and I suspect that the future norms determinations won’t have such tight control either. 

         The norms are averages and they still will be; that suggests using them as rough estimates rather than exact statistics (a point worth remembering when trying to determine if students are sufficiently fluent readers). 

          Last point: your fourth-grade teachers are correct that the texts they are testing with may not be of equivalent difficulty, which makes it difficult to determine whether or not there are real gains (or losses) being made. We've known for a long time that text difficulty varies a great deal from passage to passage. Just because you take a selection from the middle of a fourth-grade textbook, doesn't mean that passage is a good representation of appropriate text difficulty. That is true even if you know the Lexile rating of the overall chapter or article that you have drawn from (since difficulty varies across text). The only ways to be sure would be to do what Hasbrouck and Tindal did--use a lot of texts and assume the average is correct; or measure the difficulty of each passage used for assessment. The use of longer texts (having kids read for 2-3 minutes instead of 1) can improve your accuracy, too.


How to Improve Reading Achievement Powerpoint
         






Sunday, October 25, 2015

Fluency Instruction for Older Kids, Really?

School Administrator Question:  Dr. Shanahan...for grades 3-5 does it make sense to use classroom time to have students partner read? If our ultimate goal is improving silent reading comprehension, I wonder at this age level if we are not using time efficiently.

Shanahan's response.:
I get this question a lot. Since our kids are going to be tested on their silent reading comprehension, why should we bother to have them practice oral reading? The purpose quite simply is that oral reading practice has been found to have a positive impact on students’ silent reading comprehension. The National Reading Panel reviewed 16 experimental studies in which students practiced their oral reading with a partner (e.g., parents, teachers, other students, and in one case, a computer), with rereading (they should be reading texts that are relatively hard, not ones they can read fluently on a first attempt), and with feedback (someone who helps them when they make mistakes). In 15 or the 16 studies, the kids who were engaged in this kind of activity ended up outperforming the control students in silent reading comprehension. There have been many additional studies since that time—across a variety of ages, with similar results.

Although oral reading practice improves oral reading that isn’t the reason we do it. We want students to practice making the text sound meaningful—which means reading the authors’ words accurately, reading with sufficient speed (the speed of language—not hurrying or racing through a text), and with proper expression or prosody (putting the pauses in the right places, making the text understandable to English speakers). As they learn to do that with increasingly complex texts, their ability to do that with silent reading improves.

Teachers are often told to stop this in the primary grades, and the Common Core standards only include fluency teaching through grade 5, but by 8th grade, oral reading fluency differences still explain 25% of the variance in reading comprehension. In other words, if you could make all the 13-year-olds equal in fluency, you’d reduce the comprehension differences by 25%.

It’s not an either/or, of course, I prescribe both fluency instruction and comprehension instruction and the latter would definitely include silent reading of the texts. You could also argue for additional silent reading comprehension practice in social studies and science reading. However, if you only have kids practicing their silent reading, then you are slowing kids’ progress and sacrificing achievement points.

Do as you please, but as director of reading of the Chicago Public Schools, I mandated fluency instruction at those grade levels and would do so again if I still had that responsibility.

Tuesday, September 29, 2015

Of Carts and Horses: Where Fluency Teaching Fits in the Learning to Read

Our preK-5 school has a number of struggling readers, and we were told yesterday that we should focus only on fluency and accuracy, not comprehension or vocabulary. We were also told that we really shouldn't be using our grade level reading materials or complex texts in the classroom until students are fluent and accurate. I'd love to hear your thoughts on what we do when we have large numbers of struggling readers.

I get lots of questions about the sequence of instruction. In this case the issue is fluency versus comprehension; more often it is about phonics, both about the sequence of phonics elements, or like this question, whether decoding proficiency is prerequisite to any other literacy teaching?  

Let’s face it… in life there are times when sequence… definitely put your car in gear before you step on the gas, and my grandchildren love knowing that you have to put your socks on before your shoes if you want things to work out right.

But there are also lots of times when order doesn’t really matter (unless you have obsessive-compulsive disorder, but that’s another topic altogether).

For example, when you’re eating your dinner, no one is likely to care much whether you take a bite of potatoes first or a bite of the green beans. It usually doesn’t matter whether you read the sports section first or the news. And, who cares whether you put on your right or left earring first? It makes no difference so such orderings are left to one’s discretion, comfort, or habit.

So what’s the right answer concerning whether teachers should focus only on “fluency and accuracy” before comprehension and vocabulary?

I think whoever is telling teachers that they need to accomplish oral reading fluency before comprehension is wrong. This notion shows a weak understanding of the oral reading fluency concept and what it contributes to literacy learning.

Fluency is not a single skill as much as an amalgamation of skills. It has three dimensions, not two (it is more than just accuracy and speed, but also includes making the oral reading sound meaningful—expression or prosody). Students both develop decoding and comprehension skills through fluency practice, but they also learn to incorporate those skills within their oral reading (how would one know what to do with the homographs—like minute, digest, resent if comprehension isn’t part of it?).

Of course, if contextual information isn’t entering the system, then students’ fluency development will lag. If it is lagging in the first place (which sounds like the case here), then extra fluency practice is sensible… but if decoding and comprehension instruction is being delayed until fluency is developed, then where do they get the skills and knowledge that is part of what makes fluency go?

If the question had been about whether one should wait to work on fluency and comprehension until decoding was accomplished up to some criterion, I would be giving a similar answer. Decoding is central to beginning reading instruction and I don’t believe that we should stint on it. However, that doesn’t mean teachers shouldn’t, alongside, be emphasizing comprehension (initially listening, eventually reading), oral reading fluency, and vocabulary, too.

And, no, there is no particular sequence of phonics that needs to be taught—though a planful sequence is important (it just doesn’t matter whether you teach a “d” first or an “m”). Similarly, though the line of development in fluency tends to go from accuracy to speed to prosody, you still should emphasize all of them throughout (that, "read this as fast as you can" is foolishness).

I suspect some of the confusion over this comes from a misunderstanding of how the research is done on these literacy components. People get it in their heads that the phonics studies must have only taught phonics. That tends not to be the way these studies are done. In fact, the most typical experimental design has been that the experimental and control subjects both get a fairly comprehensive instructional program, but the experimental group gets an enhanced, special, super-duper version of whatever the component of interest may be (e.g., vocabulary, phonics, fluency, comprehension strategies). That often means that both groups receive some phonics or some fluency work depending on the individual teachers, but that the experimental ones would be more likely to teach these skills more thoroughly or extensively.

We may be thinking that this is the design:
      Experimental Group                                 Control Group                                                 
         Fluency Instruction                                     No Fluency Instruction

But it is more likely to like this:
           
 Experimental Group                                         Control Group
   Daily XYZ Reading Program Instruction          Daily XYZ Reading Program Instruction
         +Fluency Instruction                                       + Nothing

And what that means is that it wasn’t the additional fluency or phonics that was raising reading achievement, but that additional instruction was effective when added to an ongoing comprehensive program of teaching. 

One of the things that may be making fluency instruction work is that kids are daily learning about letters, sounds, and spelling patterns—and without that information, the fluency teaching on its own might not help as much. Similarly, the work being done to build students’ knowledge of language, content, and comprehension may also be contributing to children’s fluency growth.

As proposed here, cutting kids off from such simultaneous opportunities to learn may both slow their progress in developing fluency and may make fluency more of a parlor trick than a dynamic part of the reading process involving the coordination of high speed decoding with the context of language and ideas.