Lynn S. Fuchs, Douglas Fuchs, Michelle K. Hosp and Joseph R. Jenkins' article, Oral Reading Fluency as an Indicator of Reading Competence: A Theoretical, Empirical, and Historical Analysis, from Scientific Studies of Reading, 5(3), 2001, has been referenced in several of the articles I have recently read about fluency so I worked to dig up a copy of it. This article examines if oral reading fluency as defined by correct words per minute (CWPM) read predicts reading competence. The short answer is yes.
The authors discuss the theoretical foundation of this concept. Although researchers have different theories on how they arrived at the idea, they concur that being able to read fluently frees up mental processing capacity for comprehension skills. They refer to couple of 1975 studies by Posner and Snyder which claims two independent processes are at work during reading- one automatic and one optional- 1) memory location is accessed and related semantic memories are triggered and 2) context is allowed to trigger prediction about the upcoming word. Good readers rely on the automatic process whereas poor readers rely on a balance of the two. This could explain why our struggling readers use clues to the word such as initial consonant to "read" the word and make a mistake and then go on to fabricate the rest of the sentence in a meaningful manner. The poor readers do not have effective mechanisms for automatically accessing semantic information and consequently create mental predictions that lead to misreading and/or misunderstanding.
The authors note that text fluency and list fluency account for 70% of the variance on Iowa test scores. Since oral fluency so accurately reflects reading comprehension, it is odd that we do not focus more on it. I think part of the challenge is how complex and interrelated it with other aspects of reading fluency is. You will not be fluent if you do not have phonemic awareness and phonetic decoding skills. Furthermore reading fluently is extremely difficult if your vocabulary skills do not match the vocabulary of the passage. When this is added to the emphasis that without comprehension there is no reading, it may appear that if you attend to these other skills, fluency will automatically follow. This is not true. I have worked with language delayed students who can read beautifully; they have excellent decoding and good awareness of punctuation during reading- but they do not understand what they read. More often however, I have worked with students for whom reading is a word by word process. If they are read the material they get it, but if they need to process the printed word they are at a loss. We need to address the entire Parthenon of reading components in order to produce successful readers.
One interesting thing the authors note is that oral reading had far better correlations with reading comprehension than silent reading. The authors postulate that perhaps the self-recording of progress in silent reading is overstated. I wonder if reading aloud offers more comprehension clues. Listening to yourself read could help you understand. It might lessen the likelihood that skipped words or lines occur and it might increase the recognition of vocabulary. In some ways it is easier to self-monitor oral reading than silent reading. We suggest sub-vocalization for struggling readers and for proofreading. We do this because it supports comprehension on some level.
While the authors recommend that fluency be a part of every reading lesson, they do not endorse reading for prosody. This is because there is tremendous variation in prosody assessments to the point of their being mostly unreliable and invalid. In light of the indications that prosody reflects some comprehension I am curious regarding this restriction. I have heard very competent readers read dreadfully (my Shakespeare professor with a stutter who read in a monotone) and mediocre readers read with wonderful emphasis (my struggling reader who read a favorite Silverstein poem). I believe that prosody has a role in fluency, but perhaps we overemphasize this component when it comes to documenting student reading success.