SCIENTIFIC STUDIES OF READING, 5(3), 203–210 Copyright © 2001, Lawrence Erlbaum Associates, Inc.
Introduction to This Special Issue: The DNA of Reading Fluency Edward J. Kame’enui and Deborah C. Simmons University of Oregon
In The Double Helix, James Watson (1968) recounted with compelling candor and uncanny insight that “frightening and beautiful experience” (Feynman, as cited in Watson, 1968, p. i) of making a truly scientific discovery. Of course, the discovery Watson and his colleague Francis Crick made nearly 50 years ago was the double-helical structure of DNA. For many readers, the intrigue of Watson’s personal account is not necessarily in his discovery that four letters compose the entire alphabet of heredity, spelled out in a unique morphological arrangement of nucleotides bonded to sugar–phosphate bases. Instead, what is particularly engaging about Watson’s story is the genuine process of dubitation that he and the international community of scientists (e.g., chemists, physicists, biologists in England, France, and the United States) unwittingly revealed about the nature of scientific inquiry, particularly the unscientific hunches that science sometimes permits when exploring nature’s mysteries. Throughout The Double Helix, Watson (1968) exposed his peculiar sensibility about the structure of DNA, such as his Keatsian notions of beauty as truth and truth as beauty. For Watson and Crick, the intrinsic elegance of the DNA structure was obvious from the start: “The idea was so simple that it had to be right. … A structure this pretty just had to exist” (Watson, 1968, p. 131). In the end, Watson and Crick’s aesthetic hunches and unwieldy model building, greeted skeptically by many prominent scientists at the time, were confirmed by bold experimental X-ray data from other labs. Fluent reading, like the “thread of life” itself (Kendrew, 1966), is intrinsically elegant in both form and cadence (and perhaps biochemical valence). We certainly know it when we see it, and we are quick to celebrate it, along with the trajectory of success it portends. Rauding theory (Carver, 1981, 1983, 1984) notwithstanding, a first grader orally reading a grade-level passage at an uncommon rate of 90 correct words per minute is an awesome sight. Likewise, we readily recognize when read-
204
KAME’ENUI AND SIMMONS
ing is not fluent but is wickedly fractured and laborious in flow and purpose, when words misfire and do not enjoy a private voice or a public audience. Anyone who has been in the presence of a child or young adult unable to read an appropriate-level passage with the words executed accurately, effortlessly, and instantly one after another with unwavering prosody understands why reading fluency is elusive and bewitching. Clearly, the ability to read accurately and rapidly is so fundamental to reading success that it just has to be right. Reading fluency, however, is not an all-or-nothing proposition, even though the result (i.e., words in text read accurately and rapidly) is often treated as a vulgar dichotomy (e.g., fluent reader vs. nonfluent reader). Instead, fluent reading is plainly developmental and represents an outcome of well-specified sublexical and lexical processes and skills developed for most children over a bounded period of pedagogical time (e.g., kindergarten to Grade 6) and in rather complex host environments known as schools (Kame’enui, Simmons, & Coyne, 2000; Simmons et al., 2000). What reading fluently precisely means is not entirely clear. Even less clear are the theoretical underpinnings and the cognitive mechanisms that best explain “the speed and effortlessness with which [readers] seem able to breeze through text” (Adams, 1990, p. 409). In this special issue, the historical, theoretical, empirical, and instructional character of reading fluency are examined, but notably within the constraints of three articles. Curiously, despite the substantial and coercive efforts in reading research over the last 3 or 4 decades (Adams, 1990; Chall, 1967; National Reading Panel, 2000; National Research Council, 1998; Stanovich, 2000), reading fluency as a construct does not enjoy definitional, theoretical, empirical, or instructional consensus in the research literature (see Wolf & Katzir-Cohen in this issue; see also Nathan & Stanovich, 1991; National Reading Panel, 2000, chap. 3, p. 6). For all intents and purposes, reading fluency is eonomine; that is, it is a term so broad and unsatisfactory in meaning that little insight and understanding are gained beyond the mere use of the term. Fluency as eonomine, for example, is revealed in the National Reading Panel’s (2000) historical analysis of the term and its application. Specifically, the National Reading Panel traced the “changing concepts of fluency” from “high-speed word recognition” fluency, which marks the release of a reader from the algorithm of lower level subskills processing implicated in decoding to fluency that extends “beyond word recognition” to “comprehension processes as well” (chap. 3, p. 6). In addition, the National Reading Panel noted that “there has been a high degree of overlap in the use of the terms such as ‘automaticity’ and ‘fluency’” (chap. 3, p. 7). The historical blurring of these terms began, in part, with the early emphasis on word recognition derived from LaBerge and Samuels’ (1985) commitment to the role of attention mechanisms in information processing and their interest in capturing “the basic principles of automaticity in perceptual and associative processing” (p. 692). Interestingly, the National Reading Panel did not resolve the definitional
INTRODUCTION
205
ambiguity between automaticity and fluency but instead offered an interesting demur that reinforced the interchangeable use of the terms. In the report, automaticity was viewed as “a continuum rather than a dichotomy,” and “the importance of thinking of fluency as a continuum” (chap. 3, p. 8) was also conceded. The National Reading Panel’s (2000) overall analysis of fluency research was unambiguous in the prominent role it assigned to the development of efficient or high-speed word recognition skills and the importance of “increased practice and repeated exposures to the words in the texts that the student reads” (chap. 3, p. 8). It appears clear and unyielding that practice and exposure to print is essential to fluency (Adams, 1990; Anderson, Wilson, & Fielding, 1988; Biemiller, 1977–1978; Nathan & Stanovich, 1991). However, the cognitive mechanisms and processes that index and service fluency appear to be unsettled theoretically and experimentally (National Reading Panel, 2000; Stanovich, 2000) even though both persuasive and empirical evidence appear to support Logan’s (1988) instance theory of automatization, which asserted There is also evidence that automaticity is acquired only in consistent task environments, as when stimuli are mapped consistently onto the same responses throughout practice. Most of the properties of automaticity develop through practice in such environments. (p. 492)
This instance theory of automatization argues against resource limitation and for automatization that is “item-based (rather than process-based)” and involves “learning specific responses for specific stimuli” (Logan, 1988, p. 494). The learning mechanism implied is one that relies on the “accumulation of separate episodic traces with experience—that produces a gradual transition from algorithmic processing to memory-based processing” (Logan, 1988, p. 493). To wit, the National Reading Panel’s quantitative research synthesis reported on the effectiveness of two major instructional approaches to fluency development (i.e., repeated oral reading practice or guided repeated oral reading practice and formal strategies that encourage students to read more independently) because “there is common agreement that fluency develops from reading practice” (chap. 3, p. 1). But exactly how does this fluency develop? Under what practice or instructional conditions and at what levels necessary to access meaning? For what readers, under what text and task conditions, and at what criterion levels of performance? Furthermore, to what extent is it necessary to develop automaticity of sublexical and lexical units prior to developing fluency with connected text? How frequently must words occur in connected text to gain automaticity over what conditions of exposure? What investment in reading fluency is needed at what grade levels and how would the emphasis on fluency shift over time? In short, the DNA of reading fluency remains uncharted territory conceptually, theoretically, experimentally, pedagogically, and instructionally. Like Watson and Crick, we know the idea has to be right, but ideation alone is insufficient to ad-
206
KAME’ENUI AND SIMMONS
vance a society committed to the scientific study of ideas, particularly in reading. Collectively, we have much work to do to understand fully the features, mechanisms, and processes unique to reading fluency. Of course, gaining clarity and agreement on the meaning of fluency is an important first step in this work. Although the authors in this special issue are not consistently conspicuous on this matter, all implicitly call for recasting, if not reconceptualizing, fluency as both a construct and an index. For example, casting fluency without accuracy ostensibly exposes fluency as an empty construct, a word in name only. Fluency as an index of sheer speed without accuracy is a reckless indicator of processing, cognitive or otherwise. Instead, fluency should always serve to index both accuracy and speed. However trite this assertion appears, in the context of schools as complex host environments, accuracy and fluency are often treated as separate indexes of reading performance and competence. Similarly, casting accuracy without fluency invariably exposes accuracy as an inflated construct, a word that constrains how we think about reading performance and competence as indicators of general human performance (see Fuchs, Fuchs, Hosp, & Jenkins in this issue). For example, a first-grade child who reads 65 correct words with five errors in 1 min on a grade-level passage at the end of the school year represents a very different kind of reader than a first-grade child who reads 65 correct words with five errors in 4 min on the same passage. Both readers are reasonably accurate. However, assessing accuracy without fluency provides an incomplete picture of reading competence and limits potentially meaningful information for designing and delivering instruction. In fact, all authors in this special issue argue for using fluency as an index of accuracy and speed and as a primary indicator of reading competence. In addition, the authors call for expanding fluency beyond a measure of text speed and accuracy, to encompass prelexical components of word recognition, such as onset–rime, phonemic segmentation, and letter–sound correspondences. As such, fluency can be extended to index the speed and accuracy not just of words but of the constituent phonologic and alphabetic elements that compose words. For example, phonemic segmentation fluency would index the correct number of phonemes that compose a consonant-vowel-consonant word (e.g., sun) segmented in 1 min and without access to a print representation of the word. This kind of fine-grain fluency index holds enormous implications for monitoring the progress of beginning readers in the early and critical algorithmic stages of reading, where the “power-function speed-up processing” (Logan, 1988, p. 495) as a result of practice can be indexed periodically using multiple alternate forms of a phonemic segmentation fluency measure (see Good, Simmons, & Kame’enui in this issue). Indeed, the simplicity and elegance of reading fluency betray its complexity. Historically and theoretically it may be more akin to the complex threads of the Gordian knot than to the helical structure of DNA (although a spiral structure of fluency in which accuracy and speed serve as the two backbones of fluent reading
INTRODUCTION
207
makes for an expressive, but transparent, metaphor). For many practitioners, researchers, and children, fluency is the Gordian knot of reading competence. Although the three articles in this special issue do not disentangle all the complex threads of reading fluency, they do provide readers a serious examination of some of the more substantive conceptual, theoretical, and psychometric issues surrounding reading fluency. In addition, the authors offer important ways of thinking differently about reading fluency. In the first article, Wolf and Katzir-Cohen grapple directly with the nature of reading fluency, arguing that “macrolevel fluency is based on automaticity of microlevel subskills and their connections” (p. 215). Of course, this argument is made the way only Wolf can make it by providing readers an alluring and instructive history of the theories and models that best represent fluency research, beginning with Cattell and Huey and ending with the National Reading Panel’s (2000) report. This historical journey sets the stage for Wolf and Katzir-Cohen’s provocative “working definition” of reading fluency and subsequent components analysis of the structure of fluency. Based on this analysis, Wolf and Katzir-Cohen conclude “that reading fluency involves every process and subskill involved in reading” (p. 220), and as such, mischief and inefficiencies are possible from multiple sources and across a continuum of processes. The authors, however, are not deterred by this perceived conspiracy of components and processes in the development of reading fluency. Instead, in the final section of their article, they offer readers a conceptual framework and a description of an experimental fluency program designed to “squarely confront the complexity and developmental changes in fluency development” (p. 230). According to Wolf and Katzir-Cohen, fluency intervention requires greater intensity and more explicit attention to the sublexical and lexical (e.g., word-reading efficiency) requirements of reading than we realize. As an example of the kind of fluency intervention necessary, the authors describe the RAVE-O program, which stands for Retrieval, Automaticity, Vocabulary, Elaboration, and Orthography. The architecture of this program appears to hold great promise for building fluency within and between phonologic, orthographic, and semantic systems for children at risk of reading difficulties. In contrast to Wolf and Katzir-Cohen, who focus on fluency writ large, L. Fuchs, D. Fuchs, Hosp, and Jenkins focus in the second article on a particular form of reading fluency well established in reading research: oral reading fluency (ORF). These authors conceptualize ORF as “the speed with which text is reproduced into spoken language” (p. 241) and “indexed (or counted) as words read correctly per minute” (p. 242). Readers unfamiliar with this construct and metric will find the authors’ theoretical, empirical, and historical analysis provocative and persuasive. Fuchs et al. reason that reading aloud fluently serves to reveal, albeit indirectly, a reader’s reading competence on a range of lexical dimensions, including word-level processing, a lexicon of accessible words, and text comprehension.
208
KAME’ENUI AND SIMMONS
In practice, a high number of words read correctly per minute, when placed in the proper developmental perspective, indicate efficient word-level processing, a robust vocabulary knowledge base, and meaningful comprehension of the text. In contrast, a low ORF rate suggests inefficient word recognition skills, a lean or impoverished vocabulary, and faulty text comprehension skills. The authors describe a range of evidence on the viability and validity of ORF when compared (a) with more “direct” measures of reading comprehension, (b) to isolated word-reading fluency, and (c) to accuracy scores. In addition, the authors summarize a database that compares silent reading fluency and ORF as “correlates of reading comprehension performance” (p. 245). They also examine the actual use of ORF measures in practice and report its limited impact. Finally, Fuchs et al. provide readers an array of strategies for using ORF as an indicator of reading competence, including a method known as curriculum-based measurement. In the third and final article, Good, Simmons, and Kame’enui offer readers another fluency metric that is predicated on a set of phonologic and alphabetic tasks deemed predictive of advanced reading processes such as reading comprehension. Good et al. describe a set of 1-min fluency measures designed to evaluate beginning readers’ ability to negotiate both the phonologic and alphabetic requirements of reading, such as onset-rime, phonemic segmentation, pseudoword reading, and ORF. They discuss the use of these measures in predicting the outcome performance of third-grade readers on a high-stakes, statewide reading achievement test. Good et al. complement the high-stakes accountability movement in schools with a prevention-oriented assessment and intervention model that links three big ideas in beginning reading (i.e., phonological awareness, alphabetic understanding, and fluency with connected text) to specific fluency-based measures. Through the use of Dynamic Indicators of Basic Early Literacy Skills, they specify and examine the utility of a measurement timeline for the acquisition of reading skills necessary to meet expectations on high-stakes measures of reading outcomes. They apply this model in a study of four cohorts of kindergarten through Grade 3 students and demonstrate vividly the linkages between children’s performance on earlier fluency measures with later fluency measures and a high-stakes Grade 3 reading outcome measure. For example, 96% of the third-grade students who achieved the benchmark goal on ORF in the spring of third grade met or exceeded expectations on the high-stakes Oregon Statewide Assessment test. In an era of high-stakes accountability, this kind of assessment system is essential for linking student performance to content standards and instruction.
ACKNOWLEDGMENTS In addition to our own review, each article in this special issue received a minimum of three external, blind peer reviews. We were enormously fortunate to benefit from
INTRODUCTION
209
the expertise and insight of the following reviewers who were extraordinarily helpful and flexible: Virginia Berninger, University of Washington; Patricia Bowers, University of Waterloo; Douglas W. Carnine, University of Oregon; Lois Dreyer, Southern Connecticut State University; Barbara Gunn, Oregon Research Institute; Jan Hasbrouck, Texas A & M University; Kathy Howe, St. Croix River Education District; Asha Jitendra, Lehigh University; Rollanda O’Connor, University of Pittsburgh; Louise Spear-Swerling, Southern Connecticut State University; Janet Spector, University of Maine; Deborah Speece, University of Maryland; and Joseph Torgesen, Florida State University.
REFERENCES Adams, M. J. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA: MIT Press. Anderson, R. C., Wilson, P. T., & Fielding, L. G. (1988). Growth in reading and how children spend their time outside of school. Reading Research Quarterly, 23, 285–303. Biemiller, A. (1977–1978). Relationships between oral reading rates for letters, words, and simple text in the development of reading achievement. Reading Research Quarterly, 13, 223–253. Carver, R. (1981). Reading comprehension and rauding theory. Springfield, IL: Thomas. Carver, R. (1983). Is reading rate constant or flexible? Reading Research Quarterly, 18, 190–215. Carver, R. (1984). Rauding theory predictions of amount comprehended under different purposes and speed reading conditions. Reading Research Quarterly, 19, 205–218. Chall, J. (1967). Learning to read: The great debate. New York: McGraw-Hill. Fuchs, L. S., Fuchs, D., Hosp, M. K., & Jenkins, J. R. (2001). Oral reading fluency as an indicator of reading competence: A theoretical, empirical, and historical analysis. Scientific Studies of Reading, 5, 241–258. Good, R. H., III, Simmons, D. C., & Kame’enui, E. J. (2001). The importance and decision-making utility of a continuum of fluency-based indicators of foundational reading skills for third-grade high-stakes outcomes. Scientific Studies of Reading, 5, 259–290. Kame’enui, E. J., Simmons, D. C., & Coyne, M. D. (2000). Schools as host environments: Toward a schoolwide reading improvement model. Annals of Dyslexia, 50, 3–21. Kendrew, J. C. (1966). The thread of life. Cambridge, MA: Harvard University Press. LaBerge, D., & Samuels, S. J. (1985). Toward a theory of automatic information processing in reading. In H. Singer & R. B. Ruddell (Eds.), Theoretical models and processes of reading (3rd ed., pp. 689–718). Newark, DE: International Reading Association. Logan, G. D. (1988). Toward an instance theory of automatization. Psychological Review, 95, 492–527. Nathan, R. G., & Stanovich, K. E. (1991). The causes and consequences of differences in reading fluency. Theory Into Practice, 30, 176–184. National Reading Panel. (2000). Report of the National Reading Panel teaching children to read: An evidence-based assessment of the scientific research literature on reading and its implications for reading instruction. Washington, DC: National Institute of Child Health and Human Development. National Research Council. (1998). Preventing reading difficulties in young children. Washington, DC: National Academy Press. Simmons, D. C., Kame’enui, E. J., Good, R. H., III, Harn, B. A., Cole, C., & Braun, D. (2000). Building, implementing, and sustaining a beginning reading model: School by school and lessons learned. Oregon School Study Council (OSSC) Bulletin, 43(3), 3–30.
210
KAME’ENUI AND SIMMONS
Stanovich, K. E. (2000). Progress in understanding reading: Scientific foundations and new frontiers. New York: Guilford. Watson, J. (1968). The double helix. New York: Signet. Wolf, M., & Katzir-Cohen, T. (2001). Reading fluency and its intervention. Scientific Studies of Reading, 5, 211–239.