Early Identification and Intervention for Young Children with Reading/Learning Disabilities
Joseph R. Jenkins, University of Washington & Rollanda E. O'Connor, University of Pittsburgh
Learning Disabilities Summit: Building a Foundation for the Future White Papers
This paper is available in alternative formats: | Download Word | Download pdf |
We can all agree that reading is one of the principal tools for understanding our humanity, for making sense of our world, for advancing the democratic ideal, and for generating personal and national prosperity. We can agree that ability to read allows us to achieve three important goals: building knowledge (e.g., learning about the physical world); acquiring information for accomplishing tasks (e.g., installing a VCR); and deriving pleasure and feeding our interests (e.g., how our favorite athletic team has fared). Lacking reading ability, our lives would be very different. They would not be as rich.
Students with reading/learning disabilities (R/LDs) face enormous challenges learning to read. Many never reach a level of reading proficiency that allows them to build knowledge, acquire information, feed their interests, or enrich their lives. In some cases, their attempts to read result in such a degree of discouragement and frustration that reading subtracts rather than adds to their lives. For students with R/LDs, their early struggles in learning to read are a harbinger of dismal educational outcomes. Overall, students with learning disabilities leave elementary school with severely deficient reading and writing skills (deBettencourt, Zigmond, & Thornton, 1989; Deshler, Schumaker, Alley, Warner, & Clark, 1982) and leave secondary school with little or no improvement in these areas (Zigmond, 1990), with many dropping out before graduation (deBettencourt & Zigmond, 1990). This is why early identification and prevention of reading difficulties is important.
This paper summarizes (a) our current understanding of the difficulties encountered by children with R/LDs as they start down the road to reading and (b) research on early identification and intervention. The focus is children in kindergarten through second grade, although research on older children is included when it informs the understanding of problems in early reading acquisition. The paper is divided into four sections: background on skilled reading and reading disability (RD); early identification of children with R/LDs; intervention research on this population; and final thoughts on intervention approaches. We also offer short lists of sensible actions for practitioners working in this field.
BACKGROUND: SKILLED READING AND READING DISABILITY
Comprehension is the immediate goal of reading. Successful reading comprehension sits atop three essential pillars: the ability to read words; the ability to comprehend language; and the ability to access background and topical knowledge relevant to specific texts. Lacking any one of these foundations, reading comprehension suffers. Having an R/LD means having trouble with one or more of the foundation skills. Reading, language skills, knowledge, and word reading ability are all mutually dependent and reciprocally related (Stanovich, 1986). Weakness anywhere in the system can spell trouble for growth in the other foundation skills, and for reading development.
Reading Comprehension and Word Reading
Students with an R/LD may have weaknesses in any of the three foundation areas. However, during the beginning stages of learning to read, the most salient characteristic of these students is difficulty in acquiring efficient word-level reading skill. Thus, this paper focuses on assessment and treatment of word-level reading problems.
Two aspects of word reading are important for comprehension: accuracy and speed. Accurate word reading is critical to reading comprehension because the meanings that readers construct from text come via the words. No words, no meaning. If individuals cannot read words accurately, their comprehension suffers. Speed of word recognition is also strongly related to reading comprehension; individuals skilled in reading comprehension can read single words faster than individuals with poor reading comprehension (Perfetti & Hogaboam, 1975). Perfetti (1985) explained this relationship in terms of verbal efficiency and the sharing of limited cognitive resources.
According to Perfetti's verbal efficiency theory, both word recognition and comprehension processes consume attentional resources, which are known to be finite. The more attentional resources consumed by lower level processes (i.e., word identification), the fewer resources available for comprehension. Individuals who develop highly efficient word identification processes release cognitive resources for constructing and integrating meaning during reading. By contrast, individuals with inefficient word-reading skill (indicated by slow word recognition) must divide their attention between word identification and comprehension, and comprehension suffers. A major difference between skilled and unskilled readers, according to verbal efficiency theory, is efficiency at word-level processing. Whereas skilled readers read words in a split second (literally) without using conscious attention, the word reading of poor readers is inaccurate, slow, or both. Poor readers' inefficient word-level processing drains the very attentional resources needed to maximize comprehension.
By the end of grade 4, when the majority of children with R/LDs have been identified, these students already demonstrate pronounced deficits in word reading relative to their more skilled peers. The magnitude of this difference is illustrated in a study by Jenkins, Fuchs, Espin, van den Broek, and Deno (2000). These researchers asked fourth-grade students to orally read a passage of third-grade difficulty. Figure 1 shows the accuracy and fluency (i.e., mean number of words read in 1 minute) by students with R/LDs and more skilled peers (i.e., classmates who had average or above scores in reading comprehension). In 1 minute of reading, skilled comprehenders read three times more words than did students with R/LD. Accuracy levels were 98% and 86%, respectively, for the skilled and R/LD groups. These kinds of results underscore how disadvantaged elementary school students with R/LD are in word reading. It is not difficult to imagine how these students' inefficient word reading might overload working memory, making it difficult for them to connect and integrate text ideas into a coherent meaning representation.
Jenkins et al. (2000) illustrated the potential ramifications of slow word reading on comprehension. Using a procedure developed by Brown and Smiley (1977), they estimated that one new idea unit was introduced approximately every six running words in their experimental passage. Because skilled readers on average read the passage at a rate of 155 words per minute, they encountered approximately 26 idea units per minute (i.e., 155/6). By contrast, the R/LD group encountered approximately 9 idea units per minute (52/6). The temporal contiguity of ideas encountered by the two reader groups was sizable (26 vs. 9 per minute), a difference that may have consequences for comprehension. Interestingly, the 155-words-per-minute rate of the skilled readers is close to the speed with which TV news anchors read the news, which may be an optimal rate for processing verbal information. Considered in light of verbal efficiency theory, the less efficient word reading of students with R/LD overloads working memory and undermines reading comprehension.
Ways to Read Words
Because word-reading accuracy and speed are important, it is worthwhile to consider some of the ways we read words. We could "read" unfamiliar words by guessing their identity from sentence contexts, but guessing hardly qualifies as reading. Moreover, guessing words from context is a notoriously unreliable process (Adams, 1990). There are better ways to read words. For example, we can read unfamiliar words by analogy, noting their similarity to a familiar word (Goswami & Bryant, 1990). Using an analogy strategy, we might read the pseudoword flad by recognizing its similarity to known words like had, mad, and sad. Alternatively, we might read the pseudoword feab using a graphophonemic conversion strategy (i.e,, decoding) to assemble a pronunciation for the word. Decoding is sometimes referred to as phonological or alphabetic reading skill (Torgesen, Wagner, & Rashotte, 1997), because it involves mapping phonemes onto appropriate letters and letter combinations.
For skilled readers, pseudowords like feab present no challenges. We can read these words very fast, in a fraction of a second. Nor are we challenged by a pseudoword like regnessem, although we read it more slowly than feab, probably because it is longer, has three syllables instead of one, and bears little resemblance to any known word. Though skilled readers can read words by analogy and decoding, most words are read by sight. For example, messenger is read much faster than regnessem even though both words have the same number of letters and syllables. In fact, regnessem is messenger spelled backwards. The difference in the time required to read regnessem versus messenger demonstrates the advantage of having stored a word-specific memory. Automatic (i.e., instant) word reading requires having words readily available in memory. The fact that skilled readers can instantly read most words tells us they have vast stores of word-specific memories (think Home Depot warehouses filled with words). Reading words by sight is sometimes referred to as orthographic reading, in contrast to phonological or alphabetic reading in which words are read by recoding or translating individual graphemes to their phonemes (Torgesen et al., 2001).
Words become sight words when their complete spellings and pronunciations are stored in memory. In the sentence, "After he delivered the package, the messenjer rode off on his bicycle," we can read the word messenjer very fast, but not as fast as messenger. Even when messenjer is embedded in a sentence, we notice something amiss in its spelling that draws our attention to this word, momentarily interrupting meaning construction. The spelling, off by only one letter, disrupts the flow of reading, because it does not match the spelling of messenger in memory.
Ehri (1998) has argued that sight words, such as messenger, are stored and remembered by their specific grapheme-phoneme connections, not just as letter strings and a pronunciation. Memory would be quickly overloaded if words were stored by spellings alone, without phonetic values. For example, if words like messenger and sword retained their pronunciations but were spelled egrsemsn and wdrso (same letters, absent their phonemic values), our sight word vocabularies, which now number in the thousands, would be severely shrunken. Fortunately instead, we are able to store and retrieve words from memory using both letters and their phonemic values. According to Ehri, skilled readers store sight words in memory on several levels. These levels include the smallest units (phoneme-grapheme linkages) as well as larger sound-spelling units (rimes, syllables, and whole words). To accomplish this feat, readers must have sufficient generic graphophonemic knowledge to allow them to represent the specific spelling-phonological connections for specific words.
Fluent reading is essentially a function of volume of sight word knowledge. The fact that skilled readers can instantly read most any word they see tells us they have vast stores of word-specific memories. The primary difference between fluent and nonfluent readers is the difference in the number of words they can read by sight.
The Basis for Orthographic (Word) Reading Skill
What's needed to establish sight words? Two requirements stand out: repetitions and decoding skill. A study by Reitsma (1983) demonstrates the importance of both repetitions and decoding skill on the development of sight words by beginning readers. By the middle of first grade, after acquiring some decoding skill, children in a laboratory study significantly decreased reading times for words they had encountered as few as four times. With additional word repetitions, reading times decreased further. These children had begun to develop word-specific memories after encountering the words only a few times.
That high frequency words like the, is, and and become sight words is not surprising, given the number of times these words are encountered in text. However, most words that we can read by sight appear only occasionally. Individuals must read extensively in order to encounter specific words often enough to allow their instant recognition (reading by sight). Anderson, Wilson, and Fielding (1988) estimated that the number of words read during independent reading by fifth-grade students ranges from 0 to more than 4 million words annually. Individuals differ greatly in their amount of independent reading; those who avoid reading encounter and learn fewer words.
A second group of children in Reitsma's (1983) study did not show savings in reading times following several encounters with the same words. These were children near the beginning of first grade. Reitsma attributed their lack of word learning to insufficient graphophonemic knowledge. They were unable to take advantage of repeated encounters with target words to form word-specific memories. It is no accident that early alphabetic reading (decoding) and orthographic reading (word identification) skills are highly related (0.70 and 0.90, respectively, in Compton, 2000; Shankweiler et al., 1999).
Studies like Reitsma's demonstrate the importance of decoding ability for word learning. Some of the most convincing evidence for the important role of decoding ability for developing sight word knowledge comes from studies of exception word reading. Exception words are those that are not strictly decodable using graphophonemic knowledge (e.g., island, yacht, sword, aisle, guide). Coltheart (1978) proposed a dual-route theory to describe how exception and regular words are processed, with regular words read through application of graphophonemic knowledge, and exception words read as visual wholes. However, research on word reading has raised doubts about dual-route theory. Decoding ability seems to be at the heart of (i.e., necessary for) learning to read both regular and exception words. Figure 2 (from Tunmer & Chapman, 1998) shows this relationship graphically. Individuals may be skilled in both exception word reading and decoding, or skilled in neither. Some students can decode well, but perform poorly on exception word reading. Presumably, they lack adequate exposure to exception words. The necessary relationship that ties decoding to exception word reading is revealed by the observation that only skilled decoders are skilled exception word readers (Gough & Walsh, 1991). Thus, decoding skill appears to be important for learning all kinds of words.
Why is decoding skill important for developing a large sight vocabulary? Or, what does the process of working out pronunciations for unfamiliar words have to do with filling mental warehouses with sight words? One hypothesis is that decoding functions as a self-teaching tool which allows children to work out the pronunciation of an unfamiliar word on several occasions, and eventually secure the once unfamiliar word in orthographic memory (Share, 1995). In effect, the capacity to decode an unfamiliar word is like having a tutor available to pronounce unfamiliar words. After children have encountered and successfully decoded an unfamiliar word on several occasions, they begin to form a word-specific memory, which results in faster word recognition, much like the first-grade children in Reitsma's study.
A second possible explanation for the strong relationship between decoding and sight word reading derives from the act of decoding itself. In assembling a pronunciation for a novel word, the reader must attend to every letter in the word, thereby bonding each letter or letter combination with the phonemes that compose the word. Forming specific connections between a word's phonemes and graphemes enables a word to be read by sight (Adams, 1990).
A third possibility is that decoding skill is a proxy for graphophonemic knowledge. If, as Ehri (1992) has proposed, skilled readers fix sight words in memory using the words' complete spellings (i.e., storing every grapheme-phoneme connection in a word), readers must have sufficient graphophonemic knowledge to allow them to represent these word-specific connections. Even in learning exception words, individuals need a minimum level of graphophonemic knowledge to exploit existing regularities (e.g., even in exception words like sword and yacht, some phonemes map to graphemes in a regular fashion). More than likely, a level of decoding skill beyond simple graphophonemic knowledge is necessary for readers to establish high-quality, word-specific orthographic memories that include multiple levels of orthographic-phonological links (involving phonemes, onsets, rimes, and syllables; Ehri, 1998; Perfetti, 1992).
A fourth possibility is that individuals who easily acquire decoding skill are the same individuals who easily remember word-specific spellings and their pronunciations. Those individuals who easily induce graphophonemic knowledge may also detect and remember the specific graphophonemic elements of newly encountered words.
The Basis for Decoding Skill
Given the necessity of decoding skill for skilled word reading, we can ask, What is necessary for decoding? Two foundation skills stand out: knowledge of spelling-sound relations (i.e., graphophonemic knowledge) and phonemic awareness. Both appear to be necessary . The essential role of graphophonemic knowledge in decoding is obvious, but that of phonemic awareness is less so. Byrne and Fielding-Barnsley (1991) found that phonemic awareness accounted for significant variance on a word-choice task, after controlling for letter-sound knowledge. Their research is consistent with the large body of research indicating that children who lack phonological awareness are likely to become poor readers (Bradley & Bryant, 1983; Fletcher et al. 1994; Juel, 1988; Share, Jorm, MacLean, & Matthews, 1984; Wagner & Torgesen, 1987; Wagner et al., 1997)
Figure 3 shows a typical relationship between phonemic segmentation skill and decoding ability at the end of grade 1 (Vadasy, 2001). Inspection of this figure reveals that phonemic segmentation skill and decoding have the same kind of necessary relationship as that observed between decoding and exception word reading. Some children performed well in both phonemic segmentation and decoding nonwords; others performed poorly on both tasks. However, only students with phonemic segmentation skill were successful in decoding, even though strong phonemic segmentation did not necessarily guarantee strong decoding.
Thus, both phonemic awareness and graphophonemic knowledge appear necessary, but not sufficient, for successful decoding. For many children, instruction in how to utilize this knowledge may also be important (Iversen and Tunmer, 1993). Fielding-Barnsley (1997) found that children who had both phonemic awareness and graphophonemic knowledge benefited from instruction that asked students to say and write the sounds within printed words as they learned to read them.
Skilled Reading and Reading Disabilities
To summarize, research on skilled reading has disclosed the following foundational skills that go into making a skilled reader: phonological awareness, graphophonemic knowledge, decoding or alphabetic reading skill, orthographic or sight reading skill, and fluency, along with language comprehension. It has also revealed the nature of the relationships among these components. Sight word reading appears necessary for maximizing reading fluency and comprehension; decoding skill appears necessary for developing a large storehouse of sight words; and knowledge of spelling-sound rules plus phonemic awareness appears necessary for alphabetic reading skill. The foundational skills are like localities along a road, where reaching distant towns depends on passing through towns on the way (Spear-Swerling & Sternberg, 1994).
What do we know about students with RD in relationship to these components of skilled reading? Research has shown that students with RD are challenged in these very areas. Their reading is not as fluent as that of skilled reader, as shown in Figure 1. Their orthographic reading skill (sight word knowledge) is substantially below that of their age-level peers (Felton & Wood, 1992). Their decoding skills are especially weak (Felton & Wood, 1992; Rack, Snowling, & Olson, 1992; Shankweiler et al., 1999). Finally, they are slow to develop phonological awareness, and their graphophonemic knowledge is less secure (Juel, 1988; Wagner, Torgesen, Laughon, Simmons, & Rashotte, 1993).
The most widely accepted view of reading disabilities traces the reading problems of young children with specific R/LDs to weaknesses in processing phonological information. This weakness includes difficulties in developing phonological awareness (Shankweiler & Liberman, 1989) as well as difficulties in accessing phonological name codes (as evidenced in slower naming speeds for known stimuli like numbers and letters; Wolf & Bowers, 1999), poorer memory for phonological stimuli (e.g., recalling a series of orally presented numbers; Torgesen, Wagner, & Rashotte, 1994), or speech perception (e.g., repeating multisyllabic nonwords; Brady, 1991). In general, students with reading disabilities have been found to perform poorly in all these areas, although not every individual with R/LD will experience difficulty in every area. These phonological processing problems surface in the earliest stages of learning to read, where children experience particular difficulty in developing alphabetic reading skills (i.e., decoding). As we noted earlier, alphabetic reading skill probably plays a prominent role in the development of orthographic (sight word) reading skill, which in turn affects the development of fluency and comprehension.
Some of the most convincing evidence that students with R/LDs have specific deficits in alphabetic reading skill comes from studies using reading-level match designs. In such studies, older (e.g., grade 4) students with R/LDs are matched with younger (e.g., grade 2) typically developing readers on orthographic reading skill (e.g., the Word Identification Word Identification subtest of the Woodcock Reading Mastery Test, or WRMT; Woodcock, 1987). Both groups are then tested for alphabetic reading skill, typically measured by performance on a nonword reading measure (e.g., WRMT, Word Attack). The logic of this design is that if students with R/LDs are merely delayed in orthographic and alphabetic reading skills, they should perform on both tasks like younger typically developing readers. However, if they have a specific deficit in alphabetic reading, their performance on the nonword measure will be significantly below that of the younger typically developing readers with whom they have been matched on orthographic reading skill.
In reviewing this literature, Rack et al. (1992) concluded that students with R/LDs demonstrate nonword reading deficits relative to their younger, reading-level control group, especially in studies that control for regression effects. Since Rack et al.'s review, similarly designed research with German-speaking children with dyslexia consistently found a specific nonword reading deficit (Landerl, Wimmer, & Frith, 1997; Wimmer, 1993, 1996). These findings are important because German is considered to be an "orthographically shallow" language in which grapheme-phoneme pronunciation rules are highly consistent, unlike English which is considered an "orthographically deep" language. Whereas the nonword deficit for English-speaking students with R/LD is reflected in reading accuracy and rate, German dyslexics are relatively accurate readers of nonwords; however, their reading speed for nonwords is greatly impaired relative to that of younger, reading-level match controls (Landerl et al., 1997).
Thus, German readers with R/LD more readily acquire decoding accuracy (although this achievement is also a struggle for them). Their relative advantage over English readers in decoding accuracy is probably the result of systematic phonics instruction along with the German language's transparent graphophonemic relationships. German readers' alphabetic reading difficulty shows up as a deficit in decoding speed. Relative to German readers, English readers with R/LDs have more difficulty in acquiring decoding accuracy, probably because the graphophonemic regularities in their reading corpus is more opaque. The important point is that across languages with alphabetic orthographies, the reading problems of students with R/LD manifest most prominently as nonword reading deficits in speed and sometimes accuracy. These reading problems appear to arise from phonological dysfunction. Reading-related phonological processing problems can be observed in the development of phonological awareness, difficulty in learning graphophonemic relations, and difficulty in using phonemic awareness and graphophonemic knowledge to decode unfamiliar words, all of which negatively affect acquisition of a sight word reading vocabulary (orthographic reading skill), fluency, and comprehension.
