Fluency as an Instructional Problem
by Richard Allington - February 26, 2007
While fluency has risen to the top of the instructional agenda there is much about fluency that seems not well understood. Problems with fluency development seems most often related to particular instructional environments and addressing these instructional factors seems the most likely route to fostering fluent reading.
Twenty-four years ago I published my first article on fluency (Arlington, 1983), noting how infrequently fluency was set as an instructional goal and how few lessons focused on fostering fluency. Today, fluency is listed as one of the five pillars of evidence-based reading instruction. This new attention and status is due, in large part, to the report of the National Reading Panel (NRP, 2000) where fluency was identified as an essential proficiency. I am happy, sort of, that fluency is finally receiving instructional attention. Sort of happy because it seems to me that a central part of my arguments about fluency, and its troublesome counterpart, dysfluency, have been largely ignored.
How instructional environments can create fluency problems
The evidence available suggests that oral reading fluency for most readers emerges with little, if any, instructional attention. That is, little attention has been paid to fostering fluency until recently and, yet, the vast majority of pupils acquired the ability to read in phrases with appropriate intonation. Or more simply, to read with expression. While the NRP identified fluent reading as a critical proficiency, most of the actual research the NRP reviewed on fluency was evaluations of remedial interventions for readers who read aloud dysfluently. The issue of why some readers fail to develop fluent reading proficiency was left largely unaddressed by the NRP. I have argued that if we consider why some readers do not or cannot read fluently we might be able to substantially reduce the numbers of pupils who exhibit fluency problems (Arlington, 2006a).
Oral reading accuracy, rate, and fluency, are interrelated (Daane, Campbell, Grigg, Goodman, & Oranje, 2005). That is, it is literally impossible to read fluently when many words are not recognized, and fluent reading is always faster than word-by-word reading. But training students to recognize words quickly does not improve fluency (Dahl & Samuels, 1977). Training students to read faster may not improve fluency (Schwanenflugel, Hamilton, Kuhn, Wisenbaker, & Stahl, 2004). Assessments that monitor only accuracy and reading rate do not provide information on the development of fluency (Arlington, 2006a). To develop fluent readers, certain instructional conditions must be present, and those conditions are often not equally available to all readers.
In my view, dysfluency (the inability to read fluently) is largely the result of placing children in problematic instructional environments. In other words, fluency development is disrupted when pupils are 1) routinely given texts that are too difficult and 2) interrupted frequently when they read. Typically, these two conditions co-occur. Pupils attempting to read a too hard text make many errors and the teacher typically feels compelled to interrupt and attempt to induce an accurate reading. Ultimately, the reader begins to read aloud more tentatively, waiting for the teacher, or other external monitor, to confirm or reject the words as he reads them. Over a longer term, the reader and the teacher develop a reciprocal response pattern where the reader hesitates, awaiting a confirmation from the teacher that the reading is correct, the teacher supplies a confirmation (e.g., right, Um huh, OK, good job) and the reader continues on and continues the same verbal dance with the teacher (Artlington, 1980). As the teacher confirms or rejects the readers responses, the hesitations become more frequent (as do the interruptions/confirmations) and the reading becomes a word-by-word reading performance (McGill-Franzen & McDermott, 1978).
Ultimately, the reader fully accepts the external monitoring and no longer self-monitors his reading activity. Thus, we can observe these pupils read the sentence A below as sentence B.
John lives in a big white house.
John lives in a big white HORSE.
This sort of substitution provides powerful evidence that the interruptive external monitoring has largely eliminated self-monitoring while reading. If you were to ask this reader, What color HORSE to you live in? He would invariably ask, Do you mean HOUSE? This pupil will never speak a sentence that confuses house/horse but with sufficient amounts of too hard reading along with interruptive monitoring we can eliminate from his reading repertoire this language self-regulation that he uses in literally every other language context.
There are two related behaviors that emerge in such interactions. First, readers begin to read words with a rising intonation pattern characteristic of a query or question. This is simply a signal to the teacher that some sort of confirmation is expected. It may also indicate that the reader knows he doesnt know whether his attempt is correct and he signals the teacher to provide feedback. Second, and often accompanying the rising intonation signal, is what I call a look-away. Here the reader simply looks up from the text, typically looking at the teacher. Again, this is an even more direct signal to the teacher to jump in and provide feedback.
All these learned behaviors mean that fluent reading is unlikely. With exposure to enough too hard text accompanied by frequent interruptions readers habituate a word-by-word reading style. They also learn to rely upon an external monitor rather than developing self-regulation abilities.
What to do
The NRP (2000) reviewed the empirical research on instructional interventions that fostered fluency development. They concluded that sufficient evidence supported the technique of repeated reading to warrant recommending it as a useful intervention strategy. In general, I concur.
But what worries me at this point in time is the sudden emergence of packaged intervention programs which provide long-term practice using the repeated reading technique as a general solution for lower-achieving readers. I worry because Ive visited intervention sites this past year where struggling readers are in their third year of using these fluency-focused programs. Three years! In my clinical experience fostering fluency in dysfluent pupils was typically accomplished in 3-6 weeks, not 3 years! Most of the fluency intervention studies reviewed by the NRP were short-term efforts, none involved a 3 year intervention design. Such an extended focus on fluency is not just unnecessary but potentially damaging.
Damaging because other studies (Kuhn, 2005) find that both vocabulary and comprehension development seem quite restricted when daily repeated readings is compared with less repeated reading practice but more independent reading. A bit of repeated reading, say one 20 minute session weekly, coupled with an added daily 20 minutes of independent reading produced comparable gains in fluency and word recognition growth to a steady diet of repeated readings (20 minutes daily). Further, reducing the time spent engaged in repeated readings and expanding the time spent engaged in independent reading produced gains in comprehension that just doing repeated reading exercises did not.
Given the evidence available and my best professional advice from clinical experiences, I recommend the following if more than a very few pupils are reading dysfluently (word-by-word with little expression).
Ensure that these pupils are practicing every day in texts they can actually read with a high-level of accuracy (98% or better).
Foster adoption of the pause-prompt-praise interaction pattern whenever pupils are reading aloud (see Arlington, 2006b for details).
Consider using one of the repeated readings techniques that involves reading with/to a fluent model (e.g., choral reading, reading with audio-tape support, echo reading) for one or two days a week for a few weeks.
Ask whether these students are engaging in sufficient amounts of high-success independent reading every day. If not add more.
Monitor oral reading fluency to see if pupils are improving. The 4-point scale that the U.S. Department of Education used in assessing fluency of 4th grade students on the NAEP reading assessment (Daane, et al, 2005) is easy to use, reliable, and free. (Available at: http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2006469)
Fluency is currently a hot topic. But much of the activity surrounding fluency development seems a bit misguided because we have, in general, failed to consider that fluency problems are most often an unintended by-product of the instruction offered the children who read dysfluently, who read typically in a word-by-word manner with little phrasing or expression. Organizing lessons for these children around repeated readings of texts may be necessary for a few weeks to ease into fluent reading. But a steady diet of repeated reading activity for weeks or months on end is neither necessary nor supported by the research reviewed by the National Reading Panel. Finally, assuring that all pupils have a daily diet of high-success reading activity may be the most powerful strategy for avoiding developing dysfluent readers.
Allington, R. L. (1980). Teacher interruption behaviors during primary grade oral reading. Journal of Educational Psychology, 72, 371-377.