The Complex Nature of Reading Fluency: AMultidimensionalView

The Complex Nature of Reading Fluency:A Multidimensional View

ROXANNE F. HUDSONUniversity of Washington, Seattle, Washington, USA

PAIGE C. PULLENUniversity of Virginia, Charlottesville, Virginia, USA

HOLLY B. LANEUniversity of Florida, Gainesville, Florida, USA

JOSEPH K. TORGESENFlorida State University, Tallahassee, Florida, USA

Reading fluency is commonly defined as reading accurately at aquick rate with appropriate prosody—a simple sounding defi-nition. In fact, this definition hides complex processes and skillsneeded to produce the seemingly effortless performance of a fluentreader. Using both theory and empirical research, the presence androle of underlying processes and knowledge such as decoding flu-ency, processing speed, vocabulary, letter sound fluency, and sightword fluency are discussed. In doing this, we explain the elementsneeded for fluent reading and how they relate to each other in amultilayered fashion in young readers, and discuss the implica-tions of this model in the development and assessment of readingfluency.

Reading fluency is a topic that has received considerable attention in recentyears. Since the Report of the National Reading Panel was published in 2000,attention has shifted from phonemic awareness and decoding to those areas

Preparation of this article was supported in part by grant #H324N040039 from the U.S.Department of Education, Office of Special Education Programs. Statements do not reflectthe position or policy of this agency, and no official endorsement by them should be inferred.

Address correspondence to Roxanne F. Hudson, University of Washington, Box 353600,Miller Hall 102, Seattle, WA 98195, USA. E-mail: [email protected]

Reading & Writing Quarterly, 25: 4–32, 2009Copyright # Taylor & Francis Group, LLCISSN: 1057-3569 print=1521-0693 onlineDOI: 10.1080/10573560802491208

4

where less consensus has been established, including reading fluency. As aresult of the renewed interest in this topic, the lack of agreement on whatreading fluency actually is has been brought to the forefront. Reading fluencyis often defined as accurate reading of connected text at a conversational ratewith appropriate prosody (e.g., Armbruster, Lehr, & Osborn, 2001; Hudson,Lane, & Pullen, 2005; National Reading Panel, 2000) and is often measured asa combination of rate and accuracy—the number of correct words read aloudin one minute (e.g., Fuchs, Fuchs, & Maxwell, 1988; Shinn, Good, Knutson,Tilly, & Collins, 1992; Torgesen, Rashotte, & Alexander, 2001). Eschewing theconcept of rate, Daane, Campbell, Grigg, Goodman, and Oranje (2005)define fluency ‘‘in terms of phrasing, adherence to the author’s syntax, andexpressiveness’’ (p. v)—in other words, prosody. Taking a slightly differencestance, Samuels (2006) defined reading fluency as ‘‘decoding and compre-hending at the same time’’ (p. 39) and suggested that rate, accuracy, andprosody are indicators that this happening.

The variability in these definitions reminds us that reading fluency isa complex, multifaceted construct. In theory, because reading fluency istypically assessed while students are reading meaningful text, one’s defi-nition of fluency might easily encompass all that is important about ‘‘pro-ficient reading’’: it is accurate and efficient, it occurs with reasonablespeed that varies with the text, and it involves good comprehension ofthe meaning of the text. The definitions outlined above differ in the extentto which they emphasize one or more of the dimensions of what is typi-cally meant by proficient reading of text. The definition offered by Daaneet al. (2005) emphasizes comprehension as the most important part of thedefinition of reading fluency (because prosody of oral reading reflectscomprehension), while the operational definition often used in studiesof oral reading fluency (e.g., Fuchs et al., 1988) emphasizes accuracyand rate of reading. Samuel’s definition (2006) is actually theoretically dri-ven, as his earlier work (LaBerge & Samuels, 1974) implies that decodingand comprehension are most able to occur together when parts of bothprocesses operate ‘‘automatically.’’ Rate is one dimension of automaticity;suggesting that decoding and comprehension must occur together in flu-ent reading is to imply that decoding, or word identification, is occurringat a reasonable rate.

When thinking about the components of a definition of reading fluency,we are reminded of this statement by Wolf and Katzir-Cohen (2001):

[T]he unsettling conclusion is that reading fluency involves every processand subskill involved in reading. . . .Unlike reading accuracy, which canbe executed without utilizing some important reading componentslike semantic processes, we argue that fluency is influenced by the devel-opment of rapid rates of processing in all the components of reading.(p. 220)

Complex Nature of Reading Fluency 5

Fuchs, Fuchs, Hosp and Jenkins (2001) add to this emphasis oncomplexity:

[O]ral reading fluency represents a complicated, multifaceted perform-ance that entails, for example, a reader’s perceptual skill at automaticallytranslating letters into coherent sound representations, unitizing thosesound components into recognizable wholes, and automatically acces-sing lexical representations, processing meaningful connections withinand between sentences, relating text meaning to prior information, andmaking inferences to supply missing information. (pp. 239!240)

These researchers, along with Adams (1990), Ehri (e.g., 1998), Labergeand Samuels (1974), and Perfetti (1985), were among the first to unpack read-ing fluency by examining the various processes, skills, and knowledgeneeded to read fluently. We join in their efforts to discern the multifacetednature of the construct in order to better understand how to assess and teachreading fluency.

MULTIDIMENSIONAL NATURE OF READING FLUENCY

The purpose of this article is to discuss our conceptual framework for readingfluency and its implications for assessment and instruction. We will explainthe elements needed for fluent reading and how they relate to each otherin a multilayered fashion, and discuss the implications of this model in theassessment and instruction of reading fluency. Before we outline our frame-work, however, we must acknowledge that we will address only part of whatmany reading scientists regard as the complete definition of reading fluency.Our focus will be primarily on the aspects of reading fluency that involveaccuracy and rate, rather than prosody or reading comprehension. We willdiscuss comprehension only as it relates to, or is facilitated by, accurate read-ing that occurs at a reasonable rate. We recognize that this considerably nar-rows the focus of this paper, but it is beyond the scope of our currentexpertise to present a framework that encompasses all aspects of fluencyin the reading process.

Theoretical Assumptions

It is important to make explicit the theoretical assumptions underlying ourconceptual framework. First, we assume that the processing taking place dur-ing reading is parallel rather than serial (McClelland & Rumelhart, 1986;Seidenberg & McClelland, 1989). Adams’s (1990) model of reading pro-cessing illustrates the interconnection of the orthography, phonology, andmorphology of words as they are read. The orthographic (visual),

6 R. F. Hudson et al.

phonological (sound), meaning, and context processors work together in aparallel fashion to decode words and derive meaning from print. Throughfeedforward and feedback information, lower-level processes affect andare affected by higher-level processes. (For a complete explanation, seeAdams, 1990.)

Our second assumption is that automaticity at the sub-lexical and lexicallevels is necessary for fluent reading. According to Logan (1988, p. 493),‘‘automaticity is memory retrieval: performance is automatic when it is basedon single step direct-access retrieval of past solutions from memory.’’ For aprocess to be automatic, it must be fast, effortless, autonomous, and ableto be completed without conscious control or attention (Laberge & Samuels,1974; Logan, 1988, 1997; Posner & Snyder, 1975). When applied to reading,these elements are easily identified. Speed can be seen in the instantaneousreading of sight words, words that are read as a whole from memory, whichis much quicker than any analytic process such as using analogy or pho-nemic decoding (Ehri, 1998). Effortlessness is obvious when observing a flu-ent reader read for hours without a break or fatigue. Effortlessness is alsolinked to lack of demand on attentional resources, meaning that when wordrecognition is automatic, attention can be devoted to understanding what isread (Laberge & Samuels, 1974). Autonomy is most easily seen by the lack ofcontrol a reader has over word recognition; the process is encapsulated andoccurs whether a reader wishes to read the words or not, as evidenced by theStroop effect (Ehri, 1987). Finally, automatic processes happen so quicklythat they are beyond conscious control or analysis by the reader. It is imposs-ible for a good reader to explain how he or she automatically reads a wordby sight.

Automaticity in reading can be thought of as a race between memory(e.g., sight word reading) and an algorithm involving analysis (e.g., use ofanalogy, context, or phonemic decoding) (Logan, 1988). When a reader canrecognize a word automatically, the memory trace will always producefaster identification of the word than a process that requires analysis andapplication of an algorithm. Automaticity is not an all-or-nothing prop-osition, but rather follows a predictable curvilinear pattern of increasingspeed to an asymptote. The main mechanism for improvement in automa-ticity is practice with consistent input (letters) and consistent output (soundand meaning) pairings (Rayner, Foorman, Perfetti, Pesetsky, & Seidenberg,2001).

Automaticity is also item-specific. Because it is based on memory traces(Logan, 1988, 1997), each letter, each rime, and each word becomes auto-matic, with little transfer to other letters, rimes, or words. However, there willbe transfer between words that share the same letter patterns (Berends &Reitsma, 2007; Ehri, 2002). This item-specific phenomenon explains theeffect text difficulty has on reading fluency because the number of words thata reader can read automatically will vary based on the difficulty and topic of


the text. No reader is either fluent or not fluent; instead we can speak of areader who is fluent in a given text.

A third assumption in our model is that automaticity in reading textfollows a predictable pattern of development, from the use of simple visualcues for words through the use of the alphabetic principle to identifyunfamiliar words in text, to reading words ‘‘at a single glance’’ usingfully-amalgamated representations in memory (Ehri, 1992). As Ehri (2005)comments:

How do children learn to read words by sight? The process at the heart ofsight word learning is a connection-forming process. . . .Readers learn sightwords by forming connections between letters in spellings and sounds inpronunciations of the words (Ehri, 1992, 1998). The connections areformed out of readers’ knowledge of the alphabetic system. (p. 170)

For a theoretical discussion of how readers develop automaticity inreading words, interested readers are directed to Ehri (1998, 2002, 2005).

Like Perfetti (1985), we assume that reading processes share limited-capacity processing resources often termed working memory. Theseresources limit the amount of processing and storage in memory that mayoccur simultaneously. According to Perfetti’s verbal efficiency theory, as pro-cesses become more efficient or automatic, they use fewer of the resources ofworking memory, allowing other processes to proceed more completely.The higher-order processes of comprehension such as proposition encoding,inferring, interpreting, and integrating information are, by their very nature,resource-intensive. In contrast, processes such as letter recognition, wordrecognition (access to the word’s name), and semantic encoding (access tothe word’s contextual meaning) may become extremely efficient and auto-matic. When these processes are sufficiently automatized, according to verbalefficiency theory, this frees up working memory space for additional, ormore complex comprehension processes. Conversely, when word recog-nition processes are not efficient, they cause a bottleneck that constrainsthe operation of comprehension processes in working memory.

Conceptual Framework

Like previous researchers, we see reading fluency as a complex orchestrationof multiple sub-processes working at different levels—letter recognition tomeaning construction—that logically should explain individual differencesin reading fluency defined in the broadest sense. We have previouslyexplained parts of our conceptual framework that focuses primarily on read-ing accuracy and rate (Hudson, Torgesen, Lane, & Turner, 2006; Torgesen &Hudson, 2006; Torgesen et al., 2001), but here we expand those beginnings,which are visually represented in Figure 1. It is immediately obvious from


Figure 1 that our discussion of the elements of reading fluency is restricted tothose that are important for accurate and rapid word recognition. What wehave explicitly left out of this model is the possible influences that compre-hension may have on reading rate and accuracy (Jenkins, Fuchs, van denBroek, Espin, & Deno, 2003b), except as that is indicated in the arrow des-cending from reading comprehension to reading fluency. We have also leftout, of course, any explicit consideration of the factors that lead to fluentcomprehension.

The terms automaticity and fluency are often used interchangeably, butthe concept of automaticity actually implies more about a response than doesthe concept of fluency. In this paper, we will restrict the use of the wordautomaticity to situations in which we mean to imply that a response is ‘‘auto-matic’ in the sense that it requires few processing resources, is obligatory,and outside of conscious control. In most cases involving practical assess-ment of fluency, we do not actually know whether the fluent response alsohas these characteristics of automaticity. We will use the terms fluency andefficiency to refer to these instances.

ELEMENTS OF DECODING EFFICIENCY

Following the model from the lowest level up to reading comprehension,we begin at the sub-processes that are related to differences in decodingautomaticity. Because beginning readers routinely encounter words they

FIGURE 1 Conceptual model of reading fluency.


have never read before while reading text at their general grade level (andthis process continues through elementary and into middle and high school),decoding unknown words quickly and accurately by identifying the soundsassociated with the letters in a word and blending these sounds together toapproximate the word’s pronunciation is an important part of our restrictedmodel of reading fluency. If any of the analytic or knowledge retrievalprocesses that are required for decoding unknown words operate slowlyor inaccurately, this should have a noticeable impact on both the speedand accuracy of decoding unknown words.

Phonemic Awareness. More than 20 years of research has established theimportance of phonemic awareness in learning to decode (e.g., NationalReading Panel, 2000; Rayner et al., 2001; Wagner & Torgesen, 1987). Decod-ing involves not only identifying the sounds associated with the letters in aword, but also blending these sounds together to form the word’s pronunci-ation. If this blending process isn’t fluent, then we see readers who have dif-ficulty blending the letter sounds they identify into a word, often leading tomultiple attempts to correctly pronounce an unknown word.

Letter Knowledge. Learning grapheme-phoneme (letter-sound) relationshipsis at the heart of the alphabetic principle. Without the knowledge of howsounds are systematically represented by letters, children cannot be success-ful readers in an alphabetic language (e.g., Adams, 1990; Ehri, 1998; NationalReading Panel, 2000). Because decoding novel words typically involvesassembling or blending together multiple letters=sounds, if these letters andthe sounds they represent are not identified automatically, then the wholedecoding process will suffer and become less efficient (Wolf & Bowers,1999). According to Adams (1990), speed and accuracy in letter recognitionis critical to the whole endeavor. If letter recognition is too slow, the activationof the first letter is gone before the last is activated, and the spelling pattern isnot perceived:

Letter identification must proceed quickly enough that the units repre-senting all of the letters within a spelling pattern are near peak excitationat once. (Adams, 1990, p. 162)

Larger Letter Patterns. Automaticity in the recognition of phonograms (i.e.,letter groups within a word that share a pattern across words such as rimesand suffixes) is also critical in the development of decoding fluency. Withoutknowledge of patterns across words, students will not be able to move tomore advanced, efficient decoding involving recognition of phonogramsrepresented by multiple letters in addition to phonemes represented by singleletters (Ehri, 2002). In addition, English is more regular at the level of rimes


(a vowel plus syllable ending) and larger chunks than at the phoneme level(Kessler & Treiman, 2003; Moats, 2000), making sound-symbol relationshipsat that level more predictable and useful in decoding words. Readers needto develop context-sensitive mappings of relationships between phonemesand graphemes as well as larger units (Brown & Deavers, 1999) to becomefluent decoders.

Evidence. We provide evidence of these proposed relationships usingresearch conducted by several of the authors. We investigated the role thesesub-processes play in explaining the decoding rate and accuracy of 209 chil-dren in second grade (Hudson, et al., 2006). To measure the decoding auto-maticity of the students, we used the Nonsense Word Fluency subtest (NWF)from the Dynamic Indicators of Basic Early Literacy Skills (DIBELS; Good &Kaminski, 2002) and the Phonemic Decoding Efficiency (PDE) subtest of theTest of Word Reading Efficiency (Torgesen, Wagner, & Rashotte, 1999). Thepredictor variables were fluency in letter sounds (LSF), phonemic blending(PBF), naming of phonograms (PGF), orthographic coding knowledge(Olson, Forsberg, Wise, & Rack, 1994), RAN-letters (CTOPP; Wagner, Torge-sen, & Rashotte, 1999), and speed of global processing (Woodcock JohnsonCognitive Battery).

We analyzed our data using structural equation modeling, and ourresults indicated that when the predictor variables were considered separ-ately, all were significantly related to decoding fluency (see Table 1 for the

TABLE 1 Predictors of Decoding Automaticity

PredictorStandardized structural

coefficient

Models with single predictorLetter sound fluency

z.45""

Phonemic blending fluencyz

.17"

Phonogram fluencyz

.93""

Orthographic choice accuracyz

.58""

RAN-lettersz !.54""

Processing speed latent variablex .25"

Model with all predictorsk

Letter sound fluency .05Phonemic blending fluency .05Phonogram fluency .85""

Orthographic knowledge .14""

RAN-letters .00Processing speed latent variable !.06"

"p< .05, ""p< .001.zModel is saturated, so no model fit statistics are available.xv2¼ .008, df¼ 1 CFI¼ 1.00 RMSEA¼ .000 (90% CI .000 to .068).kv2¼ 14.92, df¼ 11 CFI¼ .996 RMSEA¼ .042 (90% CI .000 to .091).


standardized structural coefficients). We next controlled for all of the predic-tors with a model that included all variables. In that model, processing speed,naming rate and accuracy for phonograms, and orthographic knowledge hadsignificant direct relationships to decoding fluency (see Table 1). There weresubstantial correlations among the phonological tasks (LSF, PBF, and PGF)and between each of them and RAN, in contrast to orthographic knowledge,which was not related to the phonological tasks.

For this article, we conducted a post-hoc analysis to explain relation-ships among the predictor variables and phonogram naming fluency andorthographic knowledge. These models are found in Figures 2 and 3. Ourfindings indicate that fluency in phonemic blending, letter sounds, and pho-nograms are related to each other and to decoding fluency in a complex,multilayered fashion. The effects of phonemic blending and processingspeed were mediated by automaticity in letter sounds rather than a directeffect, although other variables are also at work, as the model explains only38% of the variance in phonogram naming fluency. With orthographicknowledge, only 15% of the variance was explained, but within that vari-ance, there were no mediated or direct effects from letter sound or phonemicautomaticity, with the only effects coming from RAN-letters and processingspeed.

FIGURE 2 Direct and mediated effects structural model of relationships among predictorvariables explaining phonogram fluency. Abbreviations: WJVM¼Visual Matching subtest ofthe Woodcock-Johnson Test of Cognitive Skills, 3rd edition; WJDS¼Decision Speed subtestof the Woodcock-Johnson Test of Cognitive Skills, 3rd edition; RAN-Letters¼ Letter Namingsubtest of the Comprehensive Test of Phonological Processing. """p< .001, "p< .05.


ELEMENTS OF READING FLUENCY

Word-Related ProcessesSight word automaticity. How automatically readers can identify the

words in the passage has a large role to play in how fluently they read.We have found in previous work that the size of a reader’s sight wordvocabulary, or the proportion of words in any given passage that can berecognized by sight, plays a pivotal role in how quick and accurate a readeris (Torgesen et al., 2001), particularly for students who are below average inreading rate. By ‘‘reading by sight,’’ we mean that ‘‘sight of the written wordactivates its spelling, pronunciation, and meaning immediately in memory’’(Ehri, 1998, p. 8). If a student is asked to read a passage in which a relativelyhigh proportion of the words must be decoded analytically or identifiedby contextual inference, this will have an obvious negative effect on readingfluency.

Decoding fluency. When words are not read by sight, they must beidentified analytically. If too many words in a text must be identified analyti-cally, then overall text reading fluency will suffer. Further, it may not beenough to be able to analytically decode unknown words accurately if thatdecoding process isn’t reasonably efficient. When children are not efficient

FIGURE 3 Direct and mediated effects structural model of relationships among predictorvariables explaining orthographic knowledge. Abbreviations: WJVM¼Visual Matching sub-test of the Woodcock-Johnson Test of Cognitive Skills, 3rd ed.; WJDS¼Decision Speed subtestof the Woodcock-Johnson Test of Cognitive Skills, 3rd ed.; RAN-Letters¼ Letter Naming subtestof the Comprehensive Test of Phonological Processing. """p< .001, "p< .05.


in the use of analytic decoding processes to read unknown words, they oftenfail to use those processes when reading, and their accuracy suffers. Whenconsiderable effort is expended in deciphering a word, the reader mayexpend that same level of effort for the next one, but is unlikely to do sofor a third or fourth. It is exhausting to read text with unfamiliar words whenone’s decoding processes are not fluent.

Orthographic knowledge. How well word-specific orthographic knowl-edge is represented in a reader’s lexicon contributes to early reading achieve-ment (Ehri, 1992). Knowledge of the specific, unique visual spelling patternsin words (Vellutino, Scanlon, & Tanzman, 1994) plays a role separate fromthat of sound-symbol relationships. A commonly used method to assessorthographic knowledge developed by Olson et al. (1994) is to ask a studentwhich of two letter strings with identical pronunciations is a real word (e.g.,blume, bloom). Thomson et al. (2005) found a direct relationship betweenorthographic knowledge and reading accuracy, reading rate in connectedtext, reading comprehension, spelling, and written composition in childrenand adolescents with dyslexia. Hudson et al. (2006) also found orthographicknowledge to play a large role in the decoding fluency of second graders, afinding that suggests that naming nonsense words can occur by either usingsound-symbol relationships or analogies to larger patterns in known words.

Integration of multiple cues. Many unknown words in text cannot bedecoded completely by using phonemic decoding processes alone. Childrenmust first identify and blend individual sounds to obtain an approximate pro-nunciation for an unknown word in text, and then use their sense of themeaning of the passage to select a word that most ‘‘sounds like’’ the unknownword and makes sense in the context of the passage (Share & Stanovich, 1995;Snow, Burns, & Griffin, 1998). Integration of these graphophonic andmorphosyntactic cues is critical when determining the exact pronunciationof the word being decoded. The speed with which students can combineinformation from these multiple cues should contribute to the overall fluencyof reading connected text.

Meaning Related ProcessesReading Comprehension. There is considerable evidence to suggest

that the relationship between reading fluency (with an emphasis on accuracyand rate) and comprehension is reciprocal. Reading rate and accuracy havebeen identified as important facilitators of reading comprehension (Adams,1990; Fuchs et al., 2001) in average and disabled readers (Breznitz, 1987,1991; Chard, Vaughn, & Tyler, 2002; Dowhower, 1987). More specifically,individual differences in reading rate and accuracy in the third grade werefound in one large study to be the single most important factor in accounting


for differences in performance on a measure of comprehension of complextext (Schatschneider, Buck, Torgesen, Wagner, Hassler, Hecht, & Powell-Smith, 2004).

On the other hand, it also appears that comprehension facilitates quickand accurate reading of text. For example, words in context are read fasterthan the same words out of context (e.g., Biemiller, 1977!1978). Using aneye-movement methodology and studying children reading in German,Radach (2006) found that the major locus of development from second tofourth grade appeared to be a large reduction in the time spent re-readingpreviously fixated areas of text. Rereading text usually indicates post-lexicalsentence- and text-level processing (e.g., in order to comprehend the textand answer questions). If what we know about eye movement parametersfor adults (Radach & Kennedy, 2004) is also true for children, then the mentaleffort in the service of post-lexical comprehension is a major ingredient ofreading fluency, meaning that the more time readers spend in processingthe meaning of the text, the lower their reading fluency is in connected text.Radach concludes that ‘‘it may well make more sense to see comprehensionpredicting fluency than vice versa’’ (Radach, personal communication).Jenkins et al. (2003b) found support for the view that the relationshipbetween reading rate and comprehension is reciprocal. In examining fourthgraders, they found that reading words in context explained more variance inreading comprehension than did reading the same words in a list (70% vs.9%). They also found that the students’ reading comprehension scoreexplained more variance in oral reading rate and accuracy in connected textthan did reading the same words in a list (70% vs. 54%).

Passage context. Although passage context by itself does not play alarge role in increasing word reading fluency for skilled readers (Stanovich& Stanovich, 1995), it does provide useful support for younger andpoor readers (Ben-Dror, Pollatsek, & Scarpati, 1991; Perfetti, Goldman, &Hogaboam, 1979; Pring & Snowling, 1986; West & Stanovich, 1978). Inaddition, there is evidence of substantial priming effects on word readingdue to various types of meaning cues (e.g., Forster, 1999; Hartsuiker &Westenberg, 2000), even when the prime is presented so quickly that the par-ticipant is not consciously aware of it. Children more adept at constructingmeaning because of a larger knowledge base may experience a strongerbeneficial effect of context on reading fluency than those who are less ableto construct the meaning of a passage.

Vocabulary. It seems likely that the speed with which word meaningsare identified would also affect the rate at which a passage is read. BecausePerfetti suggests that both lexical access (word name) and semantic encoding(contextual word meaning) processes must be efficient, it is reasonableto think that reading fluency would be limited if semantic activation is not


automatic. Evidence of this possible relationship can be found in studies byPerfetti and Hogaboam (1975) and Lane et al. (2009=this issue). In addition tofinding that good comprehenders read low-frequency and nonsense wordsmore quickly than poor comprehenders, Perfetti and Hogaboam also foundthat whether a participant knew the meaning of the word significantly affec-ted the poor readers but not the good ones. When reading words they didnot know, poor comprehenders were both slower and less accurate thanwhen reading words they knew the meaning of while good readers wereequally fast and accurate with both types of words. Lane and her colleaguesfound that the receptive vocabulary of students in first and second grade pre-dicted their end-of-year decoding and reading rate and accuracy scores and,in the first grade, the slope of improvement in reading rate and accuracy inconnected text. As long as readers are under obligation to be actively think-ing about the meaning of what they are reading, speed of identification ofword meanings may play a role in limiting oral reading rate and accuracy.

Metacognition. While generally thought of as a part of reading compre-hension, we suggest that metacognitive differences are likely to also influ-ence reading rate and accuracy. While beginning readers often see readingas word recognition, better readers are likely to view reading as a prob-lem-solving activity (Walczyk, 1994). Because metacognition refers to beingaware of and regulating one’s own thinking, its application to this problem-solving is clear. The larger social context of a reading activity and readers’purposes and motivation to read a particular text will all influence how flu-ently they read. Readers make many conscious and unconscious decisionsabout how to approach a reading activity based on a wide range of factors,and these decisions are likely to affect the rate at which they read the text.

For example, the value readers set on accuracy vs. speed will affect howquickly they read. Some students may be so concerned about making errorswhen reading orally for an audience that they unnecessarily slow their rate toprovide an extra measure of insurance against mistakes. In contrast, otherstudents may place a premium on getting through the text quickly, and asa result they make more errors than they would have if they allowed them-selves to read at a little slower rate. This is an important issue to considergiven the widespread use of oral reading rate and accuracy for progressmonitoring and instructional decision making. Hudson, Torgesen, andSchatschneider (2006) looked at this phenomenon by providing differentialcues to induce second grade students to have different reading goals whilereading short narrative passages. They asked students to read carefullymatched passages either for speed or for accuracy and counted how manywords were attempted and number of errors in one minute. Hudson et al.found that, on average, students increased their reading fluency significantlyunder the speed condition while only increasing their error rate by one.Col!oon and Kranzler (2006) also studied this speed-accuracy trade off and


found that fifth-grade students read significantly faster when directed to doso than when asked to do their ‘‘best reading’’ or were given no prompt asto the reading strategy. However, in contrast to Hudson et al., students in thisstudy also made significantly more errors in the speed condition than theother two. Pressley, Hilden, and Shankland (2006) also examined the effectof varying instructions on the reading rate and accuracy of third graders read-ing three DIBELS ORF passages. They contrasted directions that provided anon-specific cue (‘‘do your best reading’), encouraged speed (read ‘‘asquickly as possible’’), or encouraged reading for comprehension (‘‘read inorder to understand the story’’). They found that for two of the passages, stu-dents read significantly faster when told to do their best reading than whencued to read for understanding. No other significant differences were found.Taken as a whole, these findings suggest that children can modify their read-ing rate with varying changes in accuracy, but do not always do so, and ask-ing students to read for understanding may encourage a slower reading rate.More research is needed on the effects of various reading goals on the read-ing of students in elementary school.

Global Processes Related to Reading and Decoding FluencyRapid automatized naming (RAN). A large body of research has demon-

strated the relationship of RAN to reading achievement across various sam-ples of typical and atypical readers (Wolf, 1997; Wolf & Bowers, 1999).Rapid automatic naming can be thought of as a measure of lexical access(Wagner et al., 1993), a measure of speed of processing verbal information(Catts, Gillispie, Leonard, Kail, & Miller, 2002), or an index of a precise timingmechanism in reading (Wolf, Bowers, & Biddle, 2000). A timing deficit inRAN has been found in poor readers as compared to good readers (Cattset al., 2002; Wolf et al., 2000), and Thomson et al. (2005) found a directrelationship between rapid naming and reading rate in connected text, butnot to accuracy measures, in children and adolescents with dyslexia.

Perfetti (1985) provides information that may help understand how RANis related to decoding and reading rate. He suggests that a reader’s rate of‘‘general symbol activation and retrieval’’ will set overall processing ratelimits so that ‘‘no other process can occur faster than the access and retrievalof an over learned symbol name’’ (p. 169). It may be that RAN-letters (as ameasure of an overlearned symbol name) is a valid measure of this basic pro-cessing rate, and provides an estimate of the overall limit that plays a roleonce the other processes have reached their maximum levels of efficiency.

Global processing speed. Because the various levels of reading sharelimited resources, it stands to reason that how quickly and completely readerscan process information will explain many aspects of reading fluency. Theprocessing speed subtests of the Woodcock-Johnson Cognitive Battery(Cross-Out, Decision Speed, and Visual Matching) require participants to hold


information in their memory while simultaneously searching for items thatmeet certain requirements (are most alike or match), make a decision aboutthe items, and count the number that should be marked. They are non-verbaland confound motor ability with pure processing speed, but provide a goodproxy for the complex, simultaneous processes of reading fluently. Interest-ingly, the subtests are not equally related to decoding and reading rate andaccuracy. In a sample of second graders, we found that decision speed wasnot significantly correlated to two measures of decoding fluency while visualmatching, which involves digit strings, was (Hudson et al., 2006). Both weresignificantly related to oral reading fluency, but visual matching had a muchstronger relationship (.45 vs. .22) (unpublished data).

Summary

The foregoing analyses indicate that effortless, fluent reading is the result ofa large number of sub-processes that must be accomplished efficiently andautomatically and that interact with each other (Breznitz, 2006). Withoutautomatic access to letter-sound relationships, quick and accurate operation ofphonemic analysis and blending processes, automatic access to knowledge ofphonograms, a large number of words that can be recognized ‘‘by sight,’’ quickaccess to vocabulary knowledge, and efficient operation of basic informationprocesses, reading fluency (at least the component of fluency involving readingrate) in reading text will suffer. We turn now to a discussion of the implicationsof this framework for assessment and instruction.

IMPLICATIONS FOR ASSESSMENT AND INSTRUCTION

The framework that we have proposed suggests that reading rate and accu-racy is the result of efficient operation of multiple processes that must be exe-cuted either simultaneously or within a very brief time span. An individualmust build fluency and automaticity at each layer to be a skilled reader.The multilayered fluency framework that we have proposed, in conjunctionwith models of development of word recognition skills such as thoseproposed by Ehri and her colleagues (e.g., Ehri, 2005), have implicationsfor both assessment and instruction.

Often, assessments of students’ reading achievement, particularly forstudents beyond the primary grades, focus on reading comprehension(e.g., Biancarosa & Snow, 2004; Durkin, 1978). Most, if not all, state account-ability tests require students to read extended passages and answer questionsor write extended responses. Based on a student’s performance on thesemeasures, it is determined that the student is either a good or poor reader.However, if the student does not perform well, this type of assessment doesnot provide information about where a breakdown is occurring (i.e., the


cause of the reading failure). Research by Buly and Valencia (2002) suggeststhat there are multiple areas of reading difficulty that can be associated with afailing score on a state test. The framework outlined in this paper suggeststhat it may be useful to assess multiple layers of fluency, in addition to otherdimensions of skill and knowledge that contribute to reading comprehen-sion, in order to obtain a more complete understanding to guide targetedreading interventions. Figure 4 illustrates our multilayered framework in ahierarchical fashion. Although the acquisition of these skills is not a perfecthierarchy, this model provides a foundation on which to base assessmentand instruction. The interrelated sub-processes from Figure 1 are depictedon a ladder. Each rung on the ladder represents a sub-lexical or lexicalprocess moving from phonemic awareness upward to the top rung whichrepresents reading comprehension. In the following sections, assessmentof efficiency or fluency in each layer will be discussed.

FIGURE 4 Multileveled framework for assessing processes and sub-processes of readingfluency.


ASSESSING ORAL READING FLUENCY

Like the definition of reading fluency, how to best assess it is currently asubject of controversy (e.g., Samuels, 2006). We suggest that to assessreading fluency, teachers must listen to students read aloud and considera student’s word reading accuracy, reading rate, and prosody. Just as defi-nitions or assessment procedures that address only reading rate and accu-racy miss critical aspects of the broader conceptualization of readingfluency, so do definitions and assessments that emphasize only prosodyand phrasing. Few would argue that although rate is not the only impor-tant aspect of fluency (broadly defined), it is one of the critical features ofthe concept.

The most common method for assessing reading fluency, frequentlyreferred to as Curriculum-Based Measurement Oral Reading Fluency(R-CBM), measures the number of words read correctly in one minute. Thismethod of assessing oral reading fluency has rich evidence of its validity andreliability (e.g., Deno, Marston, Shinn, & Tindal, 1983; Deno, Mirkin, &Chiang, 1982; Fuchs et al., 1988, 2001; Good, Simmons, & Kame’enui,2001; Hosp & Fuchs, 2005; Marston, 1989; Shinn et al., 1992; Stecker, Fuchs,& Fuchs, 2005, but see Rasinski, 2004; Samuels, 2006). However, someresearchers and teachers have suggested problems with a one-minute timingbecause a student may be able to sustain a rate for one minute that is not sus-tainable in a longer passage, potentially making it an inaccurate estimate ofthe student’s true fluency, or rate for longer passages.

Another issue some teachers have with R-CBM is the perception thatsome students, termed word callers, merely read the words aloud quicklyand accurately without comprehending them, thereby providing an over-estimation of the students’ reading achievement. Hamilton and Shinn(2003) investigated this phenomenon by comparing the reading fluencyand comprehension of teacher-identified word callers (WC) and similarly-fluent peers (SFP). They found that the SFPs scored significantly better thanthe WCs on all measures of reading comprehension with very large differ-ences between groups (E.S. from .92 to 1.36). They also found that the SFPswere significantly quicker and more accurate than the WCs (E.S.¼ 1.07). Itwould appear that the WCs were not reading fluently while comprehendinglittle; they had large deficits in both areas. A final concern about theshort, one-minute timing is that students are not required to decode andcomprehend at the same time. Samuels (2006) suggests that this is simply aspeed test, and that determining the CWPM on a short passage and askingcomprehension questions may be a better measure of oral reading fluency.We recommend a combination of these methods that will provide the mostaccurate and valid picture of a student’s reading fluency, if the goal is toassess the full complexity of the fluency construct.


Judging by the lack of psychometric information, measuring prosodyreliably is more difficult than measuring accuracy and rate, and consequently,it is done less frequently. However, many researchers maintain that to obtaina comprehensive analysis of oral reading fluency, prosody should be con-sidered (e.g., Mathson, Allington, & Solic, 2006; Rasinski, 2004; Zutell &Rasinski, 1991). This will be easier to accomplish when scales with demon-strated reliability that are easy to administer have been developed. In themeantime, teachers can listen to a student read a passage orally and attendto the student’s inflection, phrase boundaries, and expression using a rubricor checklist (e.g., Daane et al., 2005; Hudson et al., 2005; Pinnell et al., 1995;Zutell & Rasinski, 1991). The results from Miller and Schwanenflugel (2006)suggest that pitch change at the end of declarative and yes=no question sen-tences would be particularly important to monitor. The National Assessmentof Education Progress (NAEP) includes a measure with reasonable reliabilityand validity that provides a four-point scale evaluating a student’s phrasing ofwords, adherence to syntax, and expressiveness (Daane et al., 2005). TheNAEP Fluency Scale is provided in Daane et al.’s Fourth-grade students read-ing aloud: NAEP 2002 Special Study of Oral Reading (NCES 2006!469,Washington, DC: U.S. Government Printing Office).

Assessing Sight Words

As a result of the reciprocal nature between comprehension and reading flu-ency, words that appear in context will be read more quickly than wordsread in isolation. However, Ehri (2005) and Samuels (2006) point out thatin order to be fluent, readers must have unitized words such that that theyare recognized by sight; furthermore, this recognition must be automaticto ensure fluent reading (Torgesen et al., 2001). The lack of fluent sight wordrecognition skills appears to be a particularly salient aspect of the reading flu-ency problems of below average readers (Jenkins, Fuchs, van den Broek,Espin, & Deno, 2003a). Thus, particularly for struggling readers, it is impor-tant to assess the extent to which students can recognize relatively commonwords fluently. The Test of Word Reading Efficiency (TOWRE; Torgesen et al.,1999), Sight Word Efficiency subtest is an easy-to-administer assessment ofword reading fluency. This subtest assesses how many words a studentcan read aloud in 45 seconds: it is extremely reliable and has strong evidencefor criterion validity. It only has two forms, and while good for diagnosis, itshould not be used for monitoring progress in reading.

Assessing Phonogram Identification

Phonograms (i.e., letter groups within a word that share a pattern acrosswords) are critical for readers to learn. Without knowledge of patterns across


words, they will not be able to move to more advanced, efficient decodingusing chunks instead of phonemes (Ehri, 2002). In order to assess a student’sknowledge of common phonograms, we can construct a measure by listingcommon phonograms that are not words (for example, eed, op, um, ab, eam,arp). Lists of common phonograms are available in resources such as TheReading Teacher’s Book of Lists (Fry, Kress, & Fountoukidis, 2000), WordMatters: Teaching Phonics and Spelling in the Reading=Writing Classroom(Pinnell & Fountas, 1998), or district curriculum guides. In order to give thisassessment, follow these instructions from Hudson (2006):

Choose three phonograms to use as practice items, and type them on aseparate sheet. Put the practice items in front of the student and say, Iam going to ask you to read as many of these phonograms as you can.A phonogram is a set of letters we see a lot in words, so you may have seenthem before as a part of words you’ve read. Try to read each one like youwould in a word, but don’t make it into a real word if it isn’t one. I’llshow you how with the first one. Read the phonogram aloud. Now youtry the next one (point to the next one). If the response is correct, thensay, That’s right. Now read this one. If the response is incorrect, sayThe phonogram says_______. You try it (Pause for answer). Now read thisone. Put the list of phonograms in front of the student and follow thesame procedures as for measuring R-CBM. This will yield a score of cor-rect phonograms per minute.

While no norms exist for this new measure, it is reasonable to expect studentsto read phonograms at the same rate as single words at their grade level.

Assessing Phonemic Decoding

When assessing phonemic decoding, the essential element that is beingindexed is the students’ understanding of the alphabetic principle. Becauseof this, assessing decoding efficiency is necessarily different from assessingsight word fluency. When provided with a real word, there is no way todetermine if the student is using memorization or decoding skills. Becausenonwords represent novel combinations of letters that students are unlikelyto have encountered before, they most purely assess students’ knowledge ofsound-symbol relationships, or the alphabetic principle.

One resource for a test of fluency in decoding is the DIBELS non-sense word fluency (NWF) subtest (Good & Kaminski, 2002). This subtestcontains Vowel-Consonant (VC) and Consonant-Vowel-Consonant (CVC)combinations (e.g., ip, rop). Because the DIBELS provides multiple formsof this subtest, it can be used for progress monitoring as well as diagnosis.There is some disagreement about the adequacy of the DIBELS NWF subtestas a useful measure of the alphabetic principle for instructional decision


making (e.g., Fuchs, Fuchs, & Compton, 2004). Because students can receivea correct score for either saying the sounds in isolation or blended together,letter-sound knowledge and decoding are combined and students in tran-sition from single letter sounds to phonemic decoding may be penalizedby first saying the individual sounds and then blending the nonsense wordstogether (Angus, 2007). Fuchs et al. (2004) also found that the slope ofimprovement in word identification fluency explained significantly morevariance in oral reading rate and accuracy than DIBELS NWF did. However,Good et al. (2001) found that NWF had good classification accuracy whenpredicting performance on oral reading fluency (ORF), and Schatschneider(2006) found that NWF scores in fall of first grade had good specificitywhen identifying students who would go on to do well on a reading compre-hension measure at the end of second grade. Good, Baker, and Peyton(2009=this issue) also found that NWF slope explained a significant amountof variance in the ORF of first graders after controlling for initial NWF status,while Powell-Smith, Castillo, Hudson, and Dedrick (in review) had similarresults. They found that NWF slope explained a significant amount of vari-ance in both ORF and reading comprehension in first grade and ORF insecond grade students at low, moderate, and high risk after controlling forNWF initial status, vocabulary, and grade retention.

The TOWRE (Torgesen et al., 1999) provides another measure of pho-nemic decoding with its Phonemic Decoding Efficiency subtest. This subtestmeasures the number of nonsense words a student can pronounce correctlyin 45 seconds. Unlike the DIBELS NWF subtest, this measure extends beyondVC and CVC combinations to include complex letter patterns (e.g., clirt, drep-nort, plenador) and requires a fully-blended response to be correct; thus, thismeasure is appropriate for older as well as younger students. As with theSight Word Efficiency test, the reliability and evidence for validity are quitestrong and there are only two forms, so while good for diagnosis, it isinappropriate for use in monitoring progress.

Assessing Letter Knowledge

Letter knowledge has long been recognized as a robust predictor of readingachievement (Catts, Fey, Zhang, & Tomblin, 1999; Pullen & Justice, 2003;Scanlon & Vellutino, 1996; Whitehurst & Lonigan, 1998). The DIBELS andAIMSweb each provide a Letter Naming Fluency subtest. These measuresassess a student’s ability to quickly and accurately name upper and lowercaseletters. Assessments of letter knowledge should also include the extent towhich an individual knows the corresponding sound for a grapheme.In addition to a letter naming fluency measure, AimsWeb (EdFormation,Inc., 2006) provides a letter-sound fluency measure that can be used to mea-sure students’ automaticity in sound-symbol relationships.


Assessing Phonemic Awareness

Phonemic awareness includes the ability to detect, match, blend, segment,or otherwise manipulate the individual sounds (phonemes) in spoken lan-guage (Lane & Pullen, 2003). It is a necessary yet insufficient skill requiredto master phonemic decoding. Although rapid automatic naming tasksappear to have a stronger predictive value in kindergarten for later readingfluency, phonemic awareness is linked to later fluency outcomes and shouldbe examined when reading difficulties occur (Allor, 2002; Phillips & Torge-sen, 2006; Schatschneider, Fletcher, Francis, Carlson, & Foorman, 2004). Inparticular, assessment of phonemic awareness should focus on blendingand segmentation skills, because these are the two skills most closely linkedto decoding. Several measures of fluency in phonemic awareness are avail-able, including the DIBELS Phonemic Segmentation Fluency (PSF) andInitial Sound Fluency (ISF) subtests (Good & Kaminski, 2002) and AIMSwebPhoneme Segmentation Fluency. While these are the only measures offluency in phonemic awareness that we are aware of and have gonethrough substantial development and testing, assessors should pay carefulattention to the validity of the scores for individual students and purposeswhen using these measures. DIBELS ISF has modest concurrent (.36 and.60) and predictive (.36 and .45) correlations with other measures of earlyreading achievement (Good et al., 2001; Hintze, Ryan, & Stoner, 2003),while PSF has larger concurrent correlations ranging from .53 (Hintzeet al.) to .68 (Good et al., 2001) and predictive correlations ranging from.54 to .73 (Good et al., 2001).

Making Instructional Decisions

The assessment of each of the interrelated layers will determine where inter-vention is required. For example, if a student performs poorly on a measureof oral reading rate and accuracy, it is insufficient to begin providing inter-vention in fluency of connected text alone. If any one of the sub-skills is fail-ing, it may be difficult for the student to profit from typical interventions thatfocus on practice reading text. Intervention, therefore, should begin at thepoint where the breakdown occurs. For example, if it is determined that astudent has not developed fluency in decoding, then intervention shouldfocus on explicit and systematic instruction in the alphabetic principle andinclude ample decoding practice. With that said, however, we should alsoacknowledge that it is not clear at this point exactly how fluent decodingskills must be in order to provide adequate support for the growth of textreading fluency. We believe that inaccurate practice of text is not as effectiveas accurate practice of text (Schwanenflugel & Ruston, 2007), but it is notclear how far decoding skills must be developed before text reading practicebecomes an effective form of fluency instruction.


Wolf and Katzir-Cohen (2001) emphasize the importance of earlyinstruction that ensures the growth of accuracy before problems in fluencycan develop. They support approaches such as Ehri’s (1998), which beginwith an emphasis on the development of word reading accuracy and shiftto a focus on increasing the rate of processing. Ehri recommends that a pri-mary goal of first-grade reading instruction should be to help children reachthe full alphabetic phase of word reading. Practice at this level developsstudents’ decoding accuracy and readies children to read chunks and,eventually, whole words automatically.

One final point about the early instruction necessary to support readingrate and accuracy is raised in two relatively recent studies. One of these stu-dies (Miller & Schwanenflugel, 2006) showed that prosody is a relativelystrong indicator of whether or not a student is comprehending the text beingread. Another study already cited (Jenkins et al., 2003b) showed that studentswho have greater comprehension of what they are reading also read morefluently. They hypothesized that certain ‘‘automatic’’ comprehension pro-cesses develop when students read large amounts of text while attendingto meaning, and because of the reciprocal nature of reading fluency andcomprehension, these relatively ‘‘automatic’’ comprehension processes areinstrumental to the development of reading fluency. Thus, when teachersencourage students to read with prosody, they are actually encouraging themto attend to the meaning of what they read. Reading practice in which stu-dents are actively encouraged to attend to meaning thus builds both wordrecognition skills through repeated opportunities to practice reading individ-ual words and also comprehension skills that make word recognition easierwith resulting increases in reading fluency.

CONCLUSION

In this paper, we have attempted to outline the complex, multifaceted natureof reading fluency (with an admitted emphasis on the element of readingrate). The major lesson from this discussion is that teachers should not thinkof a single way of increasing reading fluency, but of multiple ways. Pro-fessional development for pre-service and in-service teachers needs to focuson ensuring a deep understanding of both the complex nature of reading flu-ency (Lane et al., 2009=this issue) and multiple methods to foster it in theirstudents. However, the research of Ash, Kuhn, and Walpole (2009=this issue)suggests that this professional development should be ongoing and deliber-ately focused on helping teachers learn research-based practice sufficientlydeeply to actually implement them in the classroom.

The early foundations for reading fluency are laid by attention to power-ful instruction in the alphabetic principle and the establishment of readingaccuracy. However, accurate readers who do not read large amounts of text


will not expand their sight word vocabulary sufficiently to be fluent readersby third grade, nor will their fluency continue to expand as they are requiredto read increasingly complex text after third grade. At the same time, we mustencourage students to read for meaning from the earliest stages of readinginstruction, as this supports the development of vocabulary, knowledgeabout the world, and the growth of automatic comprehension processes thatthemselves facilitate the growth of reading fluency. In short, the growth ofreading fluency, in all its multifaceted glory, is an outcome of many differentkinds of instruction and practice—it is the natural result of explicit, system-atic, and comprehensive instruction coupled with large amounts of carefullyorchestrated reading practice.

REFERENCES

Adams, M. J. (1990). Beginning to read: Thinking and learning about print.Cambridge, Mass.: MIT Press.

Allor, J. (2002). The relationships of phonemic awareness and rapid naming to read-ing development. Learning Disabilities Quarterly, 25, 47!57.

Angus, C. (2007). The effect of written word work using word boxes on the decod-ing fluency of young at-risk readers. Unpublished doctoral dissertation,Washington State University, Pullman.

Armbruster, B. B., Lehr, F., & Osborn, J. (2001). Put reading first: The research build-ing blocks for teaching children to read. Washington, DC: The Partnership forReading.

Ash, G. E., Kuhn, M. R., & Walpole, S. (2009=this issue). Analyzing ‘‘inconsistencies’’in practice: Teachers’ continued use of round robin reading. Reading & WritingQuarterly: Overcoming Learning Difficulties, 25(1), 87!103.

Ben-Dror, I., Pollatsek, A., & Scarpati, S. (1991). Word identification in isolationand in context by college dyslexic students. Brain & Language, 40(4),471!490.

Berends, I. E., & Reitsma, P. (2007). Orthographic analysis of words during fluencytraining promotes reading of new similar words. Journal of Research in Read-ing, 30, 129!139.

Biancarosa, C., & Snow, C. E. (2006). Reading next—A vision for action andresearch in middle and high school literacy: A report to Carnegie Corporationof New York (2nd ed.), Washington, DC: Alliance for Excellent Education.

Biemiller, A. (1977!1978). Relationships between oral reading rates for letters,words, and simple text in the development of reading achievement. ReadingResearch Quarterly, 13, 223!253.

Breznitz, Z. (1987). Increasing first graders’ reading accuracy and comprehension byaccelerating their reading rates. Journal of Educational Psychology, 79(3),236!242.

Breznitz, Z. (1991). The beneficial effect of accelerating reading rate on dyslexicreaders’ reading comprehension. In M. Snowling & M. Thomson (Eds.),


Dyslexia: Integrating theory and practice (pp. 235!243). London: WhurrPublishing Ltd.

Breznitz, Z. (2006). Fluency in reading: Synchronization of processes. Mahwah, NJ:Erlbaum.

Brown, G. D. A., & Deavers, R. P. (1999). Units of analysis in nonword reading:Evidence from children and adults. Journal of Experimental Psychology, 73,208!242.

Buly, M. R. & Valencia, S. W. (2002). Below the bar: Profiles of students who fail statereading assessments. Educational Evaluation and Policy Analysis, 24(3),219!239.

Catts, H. W., Fey, M. E., Zhang, X., & Tomblin, J. B. (1999). Language bases of read-ing and reading disabilities: Evidence from a longitudinal investigation. Scien-tific Studies of Reading, 3, 331!362.

Catts, H. W., Gillispie, M., Leonard, L. B., Kail, R. V., & Miller, C. A. (2002). The roleof speed of processing, rapid naming, and phonological awareness in readingachievement. Journal of Learning Disabilities, 35, 510!525.

Chard, D. J., Vaughn, S., & Tyler, B. J. (2002). A synthesis of research on effectiveinterventions for building reading fluency with elementary students with learn-ing disabilities. Journal of Learning Disabilities, 35, 386!406.

Col!oon, E. P. & Kranzler, J. H. (2006). Effect of instructions on curriculum-basedmeasurement of reading. Journal of Psychoeducational Assessment, 24(4),318!328.

Daane, M. C., Campbell, J. R., Grigg, W. S., Goodman, M. J., & Oranje, A. (2005).Fourth-grade students reading aloud: NAEP 2002 Special Study of Oral Read-ing. (NCES 2006-469). U.S. Department of Education. Institute of EducationSciences, National Center for Education Statistics. Washington, DC: U.S.Government Printing Office.

Deno, S. L., Marston, D., Shinn, M. R., & Tindal, G. (1983). Oral reading fluency: Asimple datum for scaling reading disability. Topics in Learning and LearningDisabilities, 2(4), 53!59.

Deno, S. L., Mirkin, P. K., & Chiang, B. (1982). Identifying valid measures of reading.Exceptional Children, 49, 36!45.

Dowhower, S. L. (1987). Effects of repeated reading on second-grade transitional read-ers’ fluency and comprehension. Reading Research Quarterly, 22(4), 389!406.

Durkin, D. (1978). What classroom observations reveal about reading comprehen-sion instruction. Reading Research Quarterly, 14, 481!533.

Edformation, Inc. (2006). AIMSweb Pro Complete. Eden Prairie, Minn: Author.Ehri, L. C. (1987). Learning to read and spell words. Journal of Reading Behavior, 19,

5!31Ehri, L. C. (1992). Reconceptualizing the development of sight word reading and its

relationship to recoding. In P. Gough, L. C. Ehri, & R. Treiman (Eds.), Readingacquisition (pp. 107!143). Hillsdale, NJ: Erlbaum.

Ehri, L. C. (1998). Grapheme-phoneme knowledge is essential for learning to readwords in English. In J. L. Metsala ,& L. C. Ehri (Eds.), Word recognition in begin-ning literacy (pp. 3!40). Mahwah, NJ: Erlbaum.

Ehri, L. C. (2002). Phases of acquisition in learning to read words and implicationsfor teaching. In R. Stainthorp and P. Tomlinson (Eds.) Learning and teaching


reading (pp. 7!28). London: British Journal of Educational Psychology Mono-graph Series II.

Ehri, L. C. (2005). Learning to read words: Theory, findings, and issues. ScientificStudies of Reading, 9(2), 167!188.

Forster, K. I. (1999). The microgenesis of priming effects in lexical access. Brain andLanguage, 68, 5!15.

Fry, E. B., Kress, J. E., & Fountoukidis, K. D. (2000). The reading teachers’ book oflists (4th ed.). San Francisco: Jossey-Bass.

Fuchs, L. S., Fuchs, D., & Compton, D. L. (2004). Monitoring early reading develop-ment in first grade: Word identification fluency versus nonsense word fluency.Exceptional Children, 71(1), 7!21.

Fuchs, L. S., Fuchs, D., Hosp, M. D., & Jenkins, J. (2001). Oral reading fluency as anindicator of reading competence: A theoretical, empirical, and historical analy-sis. Scientific Studies of Reading, 5, 239!259.

Fuchs, L. S., Fuchs, D., & Maxwell, L. (1988). The validity of informal readingcomprehension measures. Remedial and Special Education, 9(2), 20!28.

Good, R. H., Baker, S. K., & Peyton, J. A. (2009=this issue). Making sense of nonsenseword fluency: Determining adequate progress in early first grade reading. Read-ing & Writing Quarterly: Overcoming Learning Difficulties, 25(1), 33!56.

Good, R. H., & Kaminski, R. A. (Eds.). (2002). Dynamic indicators of basic early lit-eracy skills (6th ed.), Eugene, Ore.: Institute for the Development of EducationalAchievement.

Good, R. H., Simmons, D. C., & Kame’enui, E. J. (2001). The importance anddecision-making utility of a continuum of fluency-based indicators of founda-tional reading skills for third-grade high-stakes outcomes. Scientific Studies ofReading, 5(3), 257!288.

Hamilton, C., & Shinn, M.R. (2003). Characteristics of word callers: An investigationof the accuracy of teachers’ judgments of reading comprehension and oral read-ing skills. School Psychology Review, 32, 228!240.

Hartsuiker, R. J., & Westenberg, C. (2000). Word order priming in written and spo-ken sentences. Cognition, 75, B27!B39.

Hintze, J. M., Ryan, A. J., & Stoner, G. (2003). Concurrent validity and diagnosisaccuracy of the dynamic indicators of basic early literacy skills and the com-prehensive test of phonological processing. School Psychology Review, 32(4),541!556.

Hosp, M. K., & Fuchs, L. S. (2005). Using CBM as an indicator of decoding, wordreading, and comprehension: Do the relations change with grade? School Psy-chology Review, 34, 9!26.

Hudson, R. F. (2006). Assessment examiner manual. Tallahassee, Fla.: Florida StateUniversity.

Hudson, R. F., Lane, H. B., & Pullen, P. C. (2005). Reading fluency assessment andinstruction: What, why, and how? The Reading Teacher, 58(8), 702!714.

Hudson, R. F., Torgesen, J. K., Lane, H. B., & Turner, S. J. (2006). Predictors ofdecoding fluency: Explaining individual differences in young children. Inpreparation.

Hudson, R. F., Torgesen, J. K., & Schatschneider, C. (2006). Relative effects of speedand accuracy cues on the oral reading fluency of young children. In preparation.


Jenkins, J. R., Fuchs, L. S., van den Broek, P., Espin, C., & Deno, S. L. (2003a). Accu-racy and fluency in list and context reading of skilled and RD groups: Absoluteand relative performance levels. Learning Disabilities: Research & Practice,18 (4), 237!245.

Jenkins, J. R., Fuchs, L. S., van den Broek, P., Espin, C., & Deno, S. L. (2003b).Sources of individual differences in reading comprehension and readingfluency. Journal of Educational Psychology, 95, 719!729.

Kessler, B., & Treiman, R. (2003). Is English spelling chaotic? Misconceptionsconcerning its irregularity. Reading Psychology, 24, 267!289.

LaBerge, D., & Samuels, S. J. (1974). Toward a theory of automatic informationprocessing in reading. Cognitive Psychology, 6, 293!323.

Lane, H. B., Hudson, R. F., Leite, W. L., Kosanovich, M. L., Strout, M. T., Fenty, N. S.,et al. (2009=this issue). Teacher knowledge about reading fluency and indica-tors of students’ fluency growth in reading first schools. Reading & WritingQuarterly: Overcoming Learning Difficulties, 25(1), 57!86.

Lane, H. B., & Pullen, P. C. (2003). Phonological awareness assessment and instruc-tion: A sound beginning. Upper Saddle River, NJ: Allyn & Bacon.

Logan, G. D. (1988). Toward an instance theory of automatization. PsychologyReview, 95, 492!527.

Logan, G. D. (1997). Automaticity and reading: Perspectives from the instance theoryof automatization. Reading & Writing Quarterly: Overcoming Learning Diffi-culties, 13, 123!146.

Marston, D. B. (1989). A curriculum-based measurement approach to assessing:What is it and why do it? In M. R. Shinn (Ed.), Curriculum-based special chil-dren (pp. 18!78). New York: The Guildford Press.

Mathson, D. V., Allington, R. L., & Solic, K. L. (2006). Hijacking fluency and instruc-tionally informative assessments. In T. Rasinksi, C. Blachowics, & K. Lems(Eds.), Fluency instruction: Research-based best practices (pp. 105!122). NewYork: Guilford.

McClelland, J. L., & Rumelhart, D. E. (1986). Parallel distributed processing: Explora-tions in the microstructure of cognition (vols. 1 & 2). Cambridge, Mass.: MITPress.

Miller, J., & Schwanenflugel, P. J. (2006). Prosody of syntactically complex sentencesin the roal reading of young children. Journal of Educational Psychology, 98,839!853.

Moats, L. C. (2000). Speech to print: Language essentials for teachers. Baltimore:Paul H. Brooks Publishing Company.

National Reading Panel. (2000). Report of the National Reading Panel. Teaching chil-dren to read: An evidence-based assessment of the scientific research literatureon reading and its implications for reading instruction: Reports of the subgroups(NIH Publication No. 00-4754). Washington, DC: U.S. Government PrintingOffice.

Olson, R. K., Forsberg, H., Wise, B., & Rack, J. (1994). Measurement of word recog-nition, orthographic, and phonological skills. In G.R. Lyon (Ed.), Frames ofreference for the assessment of learning disabilities: New views on measurementissues (pp. 243!277). Baltimore, Md.: Paul H. Brookes Publishing Co.

Perfetti, C. A. (1985). Reading ability. New York: Oxford University Press.


Perfetti, C. A., Goldman, S., & Hogaboam, T. (1979). Reading skill and the identifi-cation of words in discourse context. Memory and Cognition, 7, 273!282.

Perfetti, C. A., & Hogaboam, T. (1975). Relationship between single word decodingand reading comprehension skill. Journal of Educational Psychology, 67(4),461!469.

Phillips, B. M., & Torgesen, J. K. (2006). Phonemic awareness and reading: Beyondthe growth of initial reading accuracy. In D. K. Dickinson and S. B. Neuman(Eds.), Handbook of early literacy research (vol. 2, pp. 101!112). New York:Guilford.

Pinnell, G. S., & Fountas, I. C. (1998). Word matters: Teaching phonics and spellingin the reading/writing classroom. Portsmouth, NH: Heinemann.

Pinnell, G. S., Pikulski, J., Wixson, K. K., Campbell, J. R., Gough, P. B., & Beatty, A. S.(1995). Listening to children read aloud. Research report ED 378550.Washington, D. C.: National Center for Education Statistics.

Posner, M. I., & Snyder, C. R. R. (1975). Facilitation and inhibition in the processingof signals. In P. M. Rabbitt. & S. Dornic (Eds.), Attention and performance (vol.5, pp. 669!682). San Diego, Calif.: Academic Press.

Powell-Smith, K. A., Castillo, J. M., Hudson, R. F., & Dedrick, R. (in review). Examin-ing the use of DIBELS Nonsense Word Fluency with first and second gradestudents in Reading First schools.

Pressley, M., Hilden, K., & Shankland, R. (2006). An evaluation of end-grade-3Dynamic Indicators of Basic Early Literacy Skills (DIBELS): Speed reading with-out comprehension, predicting little. East Lansing, Michigan State University,College of Education, Literacy Achievement Research Center (LARC).

Pring, L., & Snowling, M. (1986). Developmental changes in word recognition: Aninformation-processing account. Quarterly Journal of Experimental Psychology:Human Experimental Psychology, 38, 395!418.

Pullen, P. C., & Justice, L. M. (2003). Enhancing phonological awareness, print con-cept, and oral language skills in preschool children. Intervention in School &Clinic, 39, 87!98.

Radach, R. (2006, February). Comparisons of eye movements in second and fourthgraders. Paper presented at the Pacific Coast Research Conference, San Diego,Calif.

Radach, R., & Kennedy, A. (2004). Theoretical perspectives on eye movements inreading: Past controversies, current issues, and an agenda for future research.European Journal of Cognitive Psychology, 16, 3!26.

Rasinski, T. V. (2004). Assessing reading fluency. Honolulu: Pacific Resources forEducation and Learning.

Rayner, K., Foorman, B. R., Perfetti, C. A., Pesetsky, D., & Seidenberg, M. S. (2001).How psychological science informs the teaching of reading. PsychologicalScience in the Public Interest, 2(2), 31!74.

Samuels, S. J. (2006). Toward a model of reading fluency. In S. J. Samuels &A. E. Farstrup (Eds.), What research has to say about fluency instruction(pp. 24!46). Newark, Del.: International Reading Association.

Scanlon, D. M., & Vellutino, F. R. (1996). Prerequisite skills, early instruction, andsuccess in first grade reading: Selected results from a longitudinal study. MentalRetardation and Developmental Research, 2, 54!63.


Schatschneider, C. (2006, February). Reading difficulties: Classification and issuesof prediction. Paper presented at the Pacific Coast Regional Conference, SanDiego, California, USA.

Schatschneider, C., Buck, J., Torgesen, J., Wagner, R., Hassler, L., Hecht, S., et al.(2004). A multivariate study of individual differences in performance on thereading portion of the Florida comprehensive assessment test: A brief report.Tallahassee, Fla.: Florida State University, Florida Center for ReadingResearch.

Schatschneider, C., Fletcher, J. M., Francis, D. J., Carlson, C., & Foorman, B. R.(2004). Kindergarten prediction of reading skills: A longitudinal comparativeanalysis. Journal of Educational Psychology, 96, 265!282.

Schwanenflugel, P. J., & Ruston, H. P. (2007). Becoming a fluent reader: From theoryto practice. In M. Kuhn & P. J. Schwanenflugel (Eds.), Fluency in the classroom(pp. 1!16). New York: Guilford.

Seidenberg, M. S., & McClelland, J. L. (1989). A distributed, developmental model ofword recognition and naming. Psychological Review, 96, 523!568.

Share, D. L., & Stanovich, K. E. (1995). Cognitive processes in early reading develop-ment: Accommodating individual differences into a model of acquisition. Issuesin Education: Contributions from Educational Psychology, 1, 1!57.

Shinn, M. R., Good, R. H., Knutson, N., Tilly, W. D., & Collins, V. L. (1992). Curriculum-based measurement of oral reading fluency: A confirmatory analysis of itsrelation to reading. School Psychology Review, 21(3), 459!479.

Snow, C. E., Burns, M. S., & Griffin, P. (Eds.), (1998). Preventing reading difficultiesin young children. Washington, D. C.: National Academy Press.

Stanovich, K., & Stanovich, P. (1995). How research might inform the debate aboutearly reading acquisition. Journal of Research in Reading, 18(2), 87!105.

Stecker, P. M., Fuchs, L. S., & Fuchs, D. (2005). Using curriculum-based measurementto improve student achievement: Review of research. Psychology in the Schools,42, 795!819.

Thomson, J. B., Chenault, B., Abbott, R. D., Raskind, W. H., Richards, T., Aylward, E.,et al. (2005). Converging evidence for attentional influences on the ortho-graphic word form in child dyslexics. Journal of Neurolinguistics, 18(2),93!126.

Torgesen, J. K., & Hudson, R. (2006). Reading fluency: critical issues for strug-gling readers. In S. J. Samuels & A. Farstrup (Eds.), Reading fluency: The for-gotten dimension of reading success. Newark, Del.: International ReadingAssociation.

Torgesen, J. K., Rashotte, C. A., & Alexander, A. (2001). Principles of fluency instruc-tion in reading: Relationships with established empirical outcomes. In M. Wolf(Ed.), Dyslexia, fluency, and the brain (pp. 333!355). Parkton, Md.: York Press.

Torgesen, J. K., Wagner, R. K., & Rashotte, C. (1999). Test of word reading efficiency.Pro-Ed Publishing.

Vellutino, F. R., Scanlon, D. M., & Tanzman, M. S. (1994). Components of readingability: Issues and problems in operationalizing word identification, phonologi-cal coding, and orthographic coding. In G. R. Lyon (Ed.), Frames of referencefor the assessment of learning disabilities: New views on measurement issues(pp. 279!332). Baltimore: Paul H. Brookes.


Wagner, R., Torgesen, J., & Rashotte, C. (1999). Comprehensive test of phonologicalprocessing (CTOPP). Austin, TX: Pro-Ed.

Wagner, R. K., & Torgesen, J. K. (1987). The nature of phonological processing andits causal role in the acquisition of reading skills. Psychological Bulletin, 101,192!212.

Wagner, R. K., Torgesen, J. K., Laughon, P., Simmons, K., & Rashotte, C. (1993).Development of young readers’ phonological processing abilities. Journal ofEducational Psychology, 85(1), 83!103.

Walczyk, J. J. (1994). The development of verbal efficiency, metacognitive strategies,and their interplay. Educational Psychology Review, 6, 173!189.

West, R. F., & Stanovich, K. E. (1978). Automatic contextual facilitation in readers ofthree ages. Child Development, 49(3), 717!727.

Whitehurst, G. J., & Lonigan, C. J. (1998). Child development and emergent literacy.Child Development, 69, 848!872.

Wolf, M. (1997). A provisional, integrative account of phonological and naming-speed deficits in dyslexia: Implications for diagnosis and intervention. InB. Blachman (Ed.), Foundations of reading acquisition and dyslexia: Implica-tions for early intervention (pp. 67!92). Mahwah, NJ: Erlbaum.

Wolf, M., & Bowers, P.G. (1999). The double-deficit hypothesis for the developmen-tal dyslexias. Journal of Educational Psychology, 91, 415!438.

Wolf, M., Bowers, P. G., & Biddle, K. (2000). Naming-speed processes, timing, andreading: A conceptual review. Journal of Learning Disabilities, 33(4), 387!407.

Wolf, M., & Katzir-Cohen, T. (2001). Reading fluency and its intervention. Scientificstudies of reading, 5(3), 211!239.

Zutell, J., & Rasinski, T. V. (1991). Training teachers to attend to their students’ oralreading fluency. Theory into Practice, 30(3), 211!217.


The Complex Nature of Reading Fluency: AMultidimensionalView

Documents