FW: [Reading-hall-of-fame] what is fluency?

Sat Nov 7 06:42:52 GMT 2009

Two ways to measure fluency that require decoding and comprehension at the same time.    Instructions to student.  Read this out loud and when you are finished reading I have some questions I want to ask you to see how well you understood what you read.   The processing demands of this procedure require that the student decode and understand at the same time.  There actually are two commercial tests that require this kind of dual [decode and understand at the same time] demands. The Reading Fluency Indicator, now owned by the Pearson Group, and the Gray oral reading test.

There is actually another way to measure fluency.  As you know, Phil Gough contends that in word recognition words are processed letter by letter, and every additional letter in a word adds about 50 mesecs of processing time to decoding time.  Cattell, on other hand, using a convenience sample of graduate students at Wilhelm Wundt’s lab at Leipzig, Germany , found that words were processed as holistic units.  In order to see which one was right[ Gough or Cattell], we designed a method that could provide an answer.  We took animal words that varied in length from 3-6 letters and presented them by computer to Ss. The task for the Ss  was simple. They were told that  if the word on the computer screen is an animal word, press the button.  The computer measured accuracy and latency of response. The reasoning for the study was simple. If Gough  was correct then the longer six letter animal words should have longer processing times than the short three letter words.  On the other hand, if Cattell was correct then there should be no difference in processing time between the short and long words since a chunk is a chunk.  Incidentally, all the animal words were controlled for word frequency. Our Ss were from grades 2,4,6, and college. What we found was startling, in that both Gough and Cattell were each correct, only each was correct for students with different levels of reading skill.  Gough was right for students in grades 2 and 4, the so called beginning readers..  Cattell was correct for grade 6 and college. Beginning readers were processing letter by letter, and by grade 6 and college the studens were processing the words as chunks and there was no significant difference in processing time between the short and long animal words.  We have replicated this finding at least four times.  In fact, we have replicated this finding with the Chinese written code. As you know, written Chinese is not alphabetic, but the words in Chinese differ in the number of strokes used. My students from Taiwan found that the number of strokes in a Chinese word afftects  processing students of different degrees of reading skill the same way that long and short words affect students reading in English. 

One problem with the task on animal words is that there are relatively few words that can be used that are matched for word frequency. We even changed the task to embrace a much larger corpus of words. We now say to students that if the word on the computer screen is a real word, press the button as quickly as you can.  The results hold up. Non fluent readers are process in less than holistic units and the fluent readers are chunking the word.  Curmudgeonly yours, jay samuels  I am in good company. There are lots of curmudgeons in the HOF.

From: reading-hall-of-fame-bounces at lists.nottingham.ac.uk [mailto:reading-hall-of-fame-bounces at lists.nottingham.ac.uk] On Behalf Of Gay Pinnell
Sent: Friday, November 06, 2009 3:36 PM
To: Colin Harrison
Cc: reading-hall-of-fame at lists.nottingham.ac.uk; Jay Samuels
Subject: Re: [Reading-hall-of-fame] what is fluency?

Thanks everyone for the wonderful discussion.  We have been assessing (in oral reading) along  several dimensions of fluency--pausing, phrasing, intonation, word stress, rate ( not by WPM but by a qualitative judgment of "not too fast and not too slow)  along with an additional assessment of the reader's integration of these dimensions.  Then we provide explicit instruction in these dimensions including demonstration and prompting.  It makes a big difference.  Of course this instruction takes place on text that is at independent or instructional level for the student and that assumes high accuracy and satisfactory comprehension.  So fast visual processing of print is necessary but not sufficient for fluency ( as defined here).  The reader's goal is to reflect the deeper meaning of the text with the voice  .  We find reader's theater to be engaging and helpful but we also work in small group reading instruction coupled with specific instruction to help students use word parts efficiently (in combination with language syntax and meaning). It seems that smooth processing depends on all these cognitive actions and the physical actions of the eyes working together with the eyes so that the reader can be immersed in the meaning of the story or informational text.  What are some morr great readings or studies that explore the underlying role of fast processing!  

We are finding that when students try ONLY to improve WPM, especially if it is on the report card, they mumble over words, skip key and important words and just generally make a mess of it but may actually score high.

We also find that when you have readers using phrasing,  rate in general improves.

Several years ago I (with others--Karen you may want to comment)  did a study for NAEP that found reliable qualitative evaluation of fluency was highly correlated with comprehension.  

Now FAST is the goal and I am not so sure about the  relatio ship to comprehension.  What do you have if you skip every hard word?  What  can be done to turn this situation around?

GSP  

Sent from my iPhone

On Nov 6, 2009, at 5:07 AM, "Colin Harrison" <Colin.Harrison at nottingham.ac.uk> wrote:

Thanks to everyone for their posts on fluency. I agree with Jay- we need to consider word recognition and comprehension processes together, and to reflect more carefully on precisely what is happening when what we call ‘fluent reading’ is taking place. 

The data from eye-movement research are notoriously challenging to interpret, but we know two things for sure: saccades during fluent reading last about a quarter of a second, and the distance between one fixation and the next is not fixed- it is decided saccade by saccade by the brain, on the basis of on-line monitoring of the data stream and information available in the parafovea of vision about where next it would be most productive to alight. What this means is that word recognition is not just ’automatic’ (we might equally say ‘unattended’, ie without any conscious deployment of reflective attention of the sort that is needed in construing a totally unfamiliar word), it is incredibly rapid. Word recognition takes place in perhaps a twentieth of a second- 50 milliseconds, and this leaves a whole 200 milliseconds for the brain to pass information on pronunciation to the voice box (and of course not only does nearly everyone make tiny muscle movements in the voice box in silent reading, it’s very difficult to stop them doing so- but the key insight is that vocalisation follows, rather than precedes, word recognition), and for ‘comprehension’. 

The ‘comprehension’ processes that take place in this fifth of a second have to include proximal (word-level) semantic checks on word meaning and morphology,  local grammatical coherence checks, distal (phrase-level) semantic checks, attention to building up a representation of the meaning of the text at paragraph or whole text level, and integration with other information sources, which these days might include one’s own world knowledge, illustrations, graphics, tables, stored models of text structure (if you’re old enough to have been brought up as a practising member of a sect based in a compound somewhere in Illinois allied to the Church of Schema Theory), plus alignment of incoming data with the goals or purposes of that particular act of reading (skimming, summarizing, critiquing, etc.).

Many years ago, when I was doing research into children’s understanding of school texts, I did a little experiment. I took a section from a physics text book that explained (with a diagram) how a jet engine worked, together with a piece of writing from a good high school student in which she had attempted to explain in her own words how a jet engine worked, and asked teachers ‘Do you think this student has understood this passage’?  The results were fascinating. Most non-specialists felt that the girl had done an excellent job, but science teachers were split 50-50 on whether or not she had understood the text. So I went to a professor in the Engineering Department of the University, and asked him his opinion. He said ‘Not only does this girl not understand how a jet engine works, the author of the textbook doesn’t understand how a jet engine works!’ To me, this experience highlights both the fascinating challenge and the frustratingly impossibility of pinning down what we mean by ‘comprehension’. Clearly the teachers who said ‘yes’ to my question were giving credit for the girl’s making a reasonably intelligent attempt at a summary that was essentially a linguistic transformation (including some semantic superordinates and synonyms) that we normally take as evidence of ‘comprehension’. Those who said ‘no’ were the people who felt that comprehension went deeper than this, and that they lacked evidence of a deeper processing of the concepts of compression, thrust, flow and reaction, etc. I am making the point that ‘comprehension’ of science texts is no simpler to understand and monitor than comprehension of, say, Anna Karenina.

Coming back to the question: no, fluency is not more important than comprehension. For me, fluency in oral reading is reading that, by its accuracy, prosody and intonation offers suggestive evidence that the reader is comprehending the text. One might even go further and say that fluent reading is oral reading that can help a poorer reader to comprehend a text (my experience is that poor readers are not helped by crude text-to-speech software; they can’t understand it; but they are helped by reading while listening to fluently read text).  

Colin Harrison

  _____  

From: reading-hall-of-fame-bounces at lists.nottingham.ac.uk [mailto:reading-hall-of-fame-bounces at lists.nottingham.ac.uk] On Behalf Of Jay Samuels
Sent: Wednesday, November 04, 2009 2:46 PM
To: reading-hall-of-fame at lists.nottingham.ac.uk
Subject: [Reading-hall-of-fame] what is fluency?

Jan, you ended your e-mail message with the following question:  " So? >From a teacher's perspective I'd like to know what do we mean, what does research mean by 'fluency'?

Why is fluency so important in early reading? Is it more important than comprehension?

Maybe Don, these questions might be the starting point for an article.

Jan

What we mean by fluency is the heart of a heated controversy.  I feel partly responsible in that I did not do as good a job as a should have defining "fluency" in the National Reading Panel Final Report. The definition of fluency is derived from automaticity theory which states a person is automatic at a task when the person can do two difficult tasks at the same time. If a person can do two demanding tasks simultaneously, at least one of them is automatic. For example, driving a car through heavy traffic and talking to a passenger seated next to you. It is the driving that is automatic. Fluent readers can decode the text and understand it simultaneously.  Since the definition of fluency embraces simultaneous comprehension, your question about what is more important, fluency or comprehension, puts me into a position of "it depends".  In the end, to me, it is comprehension that is more important. Take the case of the struggling beginning reader who slowly works his way through a sentence decoding the words one by one, and then having decoded all the words retraces his steps over the sentence only this time to grasp its meaning. As a teacher, I am happy the student has comprehended the sentence. But, in my talks to students, I remind them of the following, "Beyond accuracy to automaticity".  Accurate word recognition followed by comprehension is a slow process and our goal as reading teachers is to go one step beyond and to develop fluent readers who can do both steps at the same time. 

Accurate word recognition, speed of decoding, prosody in oral reading, are all only what I call the indicators of fluency but they are not the sine qua non of fluency. Fluency is the ability to do two things at the same time, decode and comprehend simultaneously. Our profession is in need of tests that measure two things at the same time, decoding and comprehension. As I pointed out in a previous mailing DIBELS type tests encourage a regrettable mind set in some students. They think that reading speed is all that is important. We need to instill in students what I think of as "engagement" the need to engage comprehension and thinking processes as we read. 

But wait! Haven't I painted myself into a corner. Doesn't decoding precede meaning?  I worried about this problem until I started to get into the research on eye movements. It is only during the eye fixation pause in reading that there is uptake of information and processing. During the length of a typical fixation pause in reading, about 1/3 of a second in duration, that the two reading tasks can take place at the same time. In fact there are several models of what happens during an eye fixation pause that demonstrate 1/3 of a second is enough time to do the two tasks during a single eye fixation. 

The ability to read with fluency to me is a bit of a miracle. For about 8 million years our species has been hard wired to process spoken language, language by ear. But it has only been for about 7000 years that our species has been processing language by eye, i.e., reading. Yet, with training we can overcome some of the design flaws of the retina and read with fluency. jay   

S. Jay Samuels

Department of Educational Psychology

College of Education

The University of Minnesota

Minneapolis, MN 55455

612 625 5586

fax 612 624 8241

_______________________________________________
Reading-hall-of-fame mailing list
Reading-hall-of-fame at lists.nottingham.ac.uk
http://lists.nottingham.ac.uk/mailman/listinfo/reading-hall-of-fame

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.nottingham.ac.uk/mailman/private/reading-hall-of-fame/attachments/20091107/f6d2815b/attachment-0001.html