Consistency of Reading-Related Phonological Processes Throughout Early Childhood: Evidence From Longitudinal–Correlational and Instructional Studies
Joseph K. Torgesen
Stephen R. Burgess
Florida State University
In this chapter, we address a question of central importance in understanding individual differences in the ease with which children learn to read. We take as a starting point the widely accepted observation that the critical problem for most children with serious reading disabilities involves learning accurate and fluent word identification skills (Rack, Snowling, & Olson, 1992; Stanovich, 1988). We also take as established fact that a primary cause of variability among children in the rate at which they acquire word reading skill is variation “in the phonological component of their natural capacity for language” (Liberman, Shankweiler, & Liberman, 1989, p. 1). The question we address in this chapter is whether reading-related phonological skills display sufficient stability across the period of early reading development to qualify as a proximal cause of a learning difficulty that is as resistant to treatment as reading disabilities often prove to be.
After reviewing a large body of experimental and longitudinal–correlational research on the relationships between phonological abilities and reading growth, Share and Stanovich (1995) concluded:
In summary, there is virtually unassailable evidence that poor readers, as a group, are impaired in a very wide range of basic cognitive tasks in the phonological domain. This applies both to reading disabled children with discrepancies from IQ and to those without such discrepancies.… These deficits are consistently found to be domain-specific, longitudinally predictive, and not primarily attributable to non-phonological factors such as general intelligence, semantic or visual processing, (p. 9)
The evidence to which Share and Stanovich referred has been generated by a wide variety of simple-correlational, experimental, longitudinal–correlational, and case study research. In this chapter, we supplement this evidence by reporting data from both longitudinal–correlational research and training studies that document the stability of phonological processes across the elementary school period, and also indicate that it may be very difficult to overcome phonological processing weaknesses in some children.
The chapter is organized in four sections. First, we define and discuss the nature of the three most widely studied phonological processes that are related to reading: phonological awareness, phonological coding in working memory, and rapid access to phonological information in long-term memory. Then we report several different analyses providing evidence of the stability of individual differences in phonological processing skills throughout the elementary school period. The third section considers evidence on variability in response to training in phonological awareness, and we conclude with a discussion of the implications for diagnosis and instruction that follow from the data on stability and resistance to treatment presented in earlier sections.
THE NATURE OF READING-RELATED PHONOLOGICAL PROCESSES
Reading-related phonological processes can be defined generally as a set of mental activities or skills that are required in reading or learning to read, and that involve accessing, storing, or manipulating phonological information. The phonological processing difficulties of children with phonologically based reading disabilities can be shown on a variety of nonreading tasks that assess: awareness of the phonological structure of words in oral language (Bowey, Cain, & Ryan, 1992; Bruck, 1992; Fletcher et al., 1994); ability to represent phonological information in short-term memory (Brady, Poggie, & Rappala, 1989; Gathercole & Baddeley, 1993; Torgesen, Rashotte, Greenstein, Houck, & Portes, 1987); and rate of access to phonological information in long-term memory (Bowers & Swanson, 1991; Denckla & Rudel, 1976; Wolf, 1991). We now consider each of these skills in turn, and then present evidence about their relationships to one another in development.
Phonological Awareness
Beginning levels of phonological awareness involve sensitivity to the individual sounds in words, whereas more advanced levels involve explicit awareness of a word’s full phonemic structure as well as ability to manipulate individual phonemes within words. Phonological awareness involves a certain kind of knowledge about words—that they can be divided into segments of sound smaller than a syllable. This knowledge grows from a vague awareness that words can share common ending sounds, to a knowledge of the essential distinctive features of most of the phonemes in our language. This latter kind of knowledge allows children to repeat the individual phonemes in a word (saying /m/-/a/-/n/ when given the word man), or to form another word by adding or subtracting a phoneme from a word (saying fit when asked to say fist without the /s/ sound). This type of knowledge about language is frequently referred to as metalinguistic knowledge, because it is knowledge about words that is more explicit, or conscious, than that required when words are used in speaking or listening.
Children’s increasing knowledge of the unique characteristics of individual phonemes supports an increasing ability to notice, think about, or manipulate the individual sounds in words. In other words, increasing knowledge of the phonemic structure of words, and of phonemes themselves, supports increasingly complex processing of these sounds. For example, children in the first semester of first grade might be able to blend the phonemes /a/ and /t/ together to form the word at, but their ability to blend increasingly longer strings of sounds will improve with practice until they will be able to blend four, five, or six sounds together to form words. Similarly, children at the beginning of first grade can usually respond with the word an when asked to say man without saying the /m/ sound, but not until later in development will they be able to respond with the word drier, when asked to say driver without saying the /v/ sound.
Phonological awareness develops slowly in young children, because phonemes are not distinct from one another in the actual sounds we hear when words are spoken. For example, the word dog has three phonemes—it differs from log in the first, from dig in the second, and from dot in the third—but when it is spoken, these phonemes are merged into a single pulse of sound so that it is impossible to separate the phonemes without some articulatory distortion. Because of our natural capacity for language, it is not necessary to be consciously aware of the phonological structure of words in order to speak or understand words in oral language. The phonological processes that translate between the underlying structure of the word in the mental lexicon and its instantiation in speech operate outside of conscious awareness. However, in order to make sense of the alphabetic system we use in our written language, in order to understand that letters map to words at the phonemic level, children must become aware that words actually are composed of segments at the phonemic level.
Isabelle Liberman and her colleagues at Haskins Laboratories in New Haven, Connecticut, most completely developed the early theoretical basis for research on phonological awareness. They recognized that speech and reading both involve phonological processing, and they sought to determine why speech was so easy for most children to master with no explicit instruction, whereas at the same time reading was so hard for many children, even with explicit instruction. Their initial answer, which has been supported in subsequent research, was that reading requires an “awareness of the phonological structure of the words of the language, an awareness that must be more explicit than is ever demanded in the ordinary course of listening and responding to speech” (Liberman et al., 1989, p. 5).
Phonological Short-Term Memory
Verbal short-term memory tasks are included as measures of phonological processing ability because, when an individual is required to retain a short sequence of verbal items verbatim, these items are represented in working memory in terms of their phonological features (Baddeley, 1991). In other words, if an individual is asked to remember a brief string of digits, such as 7-4-8-9-2-6, these items are represented in memory by codes that utilize their acoustic, or phonological, features, rather than their visual or semantic features. Verbal short-term memory tasks appear to be sensitive to individual differences in the efficiency with which phonological information is represented in memory (Baddeley, 1991; Torgesen, 1995).
Children with phonological coding difficulties can be expected to have problems acquiring alphabetic reading skills because these coding difficulties make it hard to utilize knowledge of letter-sound correspondences in decoding words. Specifically, phonological coding inefficiencies make it difficult to perform the simultaneous, or rapidly sequential identification, comparison, and blending processes that are required to identify words by phonological/analytic strategies.
In fact, difficulty remembering exact sequences of verbal information over brief periods of time is one of the most frequently reported cognitive characteristics of children with severe reading disabilities (see Hulme, 1988; Jorm, 1983). The tasks most often used to demonstrate this difficulty are called memory span tasks. These tasks typically involve recalling sequences of random digits, words, or letters immediately after a single auditory or visual presentation. In the 1990s (Gathercole & Baddeley, 1993), another type of task has been used to assess phonological memory. This task involves the repetition of complex nonwords, such as morphglanome, or nonword phrases.
Rapid Automatic Naming of Verbal Material
Children’s ability to easily and rapidly access phonological information that is stored in long-term memory has typically been assessed in the reading literature by rapid automatic naming tasks. This type of task was first introduced as a way of predicting and understanding individual differences in reading ability by Martha Denckla and her colleagues (Denckla & Rudel, 1976; Rudel, Denckla, & Broman, 1978), and typically requires the child to name, as rapidly as possible, a series of 30 to 50 items (digits, colors, letters, or objects) printed on a page.
Theoretically, rapid naming tasks are linked to reading because they are thought to index the speed of processes that are intrinsically involved in word identification. In utilizing alphabetic reading skills, for example, the child must rapidly access, store, and interpret strings of phonemes represented by letters in words. If access to the phonological information represented by letters occurs rapidly and easily, the entire complex string of mental operations will be facilitated. Bowers and her colleagues (Bowers, Golden, Kennedy, & Young, 1994; Bowers & Wolf, 1993) suggested that rapid automatic naming tasks may be measuring processes that are particularly critical in the formation of orthographic, or whole-word, representations that are important in fluent word reading. They argued against viewing rapid automatic naming tasks as primarily phonological in nature, and instead they emphasized the visual and speed components of these tasks. They suggested that rapid naming tasks may assess the functioning of a “precise timing mechanism,” and that “slow letter (or digit) naming speed may signal disruption of the automatic processes which support induction of orthographic patterns, which, in turn, result in quick word recognition” (Bowers & Wolf, 1993, p. 70). Although empirical support for this view is still weak, it does represent an important alternative, or additional, conceptualization of the reasons why individual differences in performance on rapid naming tasks are causally related to variability in the growth of word reading skills.
We continue to believe that rapid automatic naming tasks assess an important dimension of phonological processing skill for three reasons. First, these tasks show a consistent pattern of significant correlations with other aspects of phonological skill (Wagner, Torgesen, Laughon, Simmons, & Rashotte, 1993; Wagner et al., 1997). Second, in our longitudinal research, rapid automatic naming skills show unique causal relationships with the development of early word reading skills, but not later word reading growth (Wagner et al., 1997). Current theories (Share & Stanovich, 1995) of reading skill acquisition suggest that phonological skills are more centrally involved in the growth of word reading skills in the early elementary grades (1–3) than the later grades (4–5). Finally, we have found no evidence in our longitudinal research that rapid automatic naming ability contributes uniquely to the development of orthographic representations of words (Torgesen, Wagner, Rashotte, Burgess, & Hecht, 1997). Although we assert that rapid automatic naming tasks do tap an important dimension of phonological skill, we also recognize that, when given in the serial naming format, these tasks are sufficiently complex that they may measure more than one processing skill important in reading.
Relations of Phonological Abilities to One Another and to Reading
Considering that all the tasks just discussed require some form of processing of the phonological features of language, it is logical to inquire whether they are sufficiently independent of one another to be thought of as measuring separate abilities, or whether they are better thought of as measures of a single underlying ability. This question was directly addressed in two studies that we reported as part of a longitudinal investigation of the relationships between phonological processing and reading.
The first (Wagner et al., 1993) was actually a cross-sectional pilot study in which we tested our measurement models for each type of phonological skill and also examined their relationships to one another. In this study, which employed 95 kindergarten and 89 second-grade children as subjects, measures of phonological awareness were further subdivided into analytic and synthetic tasks. Analytic measures required children to identify individual phonemes within whole words, whereas synthetic tasks required children to combine separately presented phonemes into words. Similarly, naming tasks were subdivided into tasks that required the continuous naming of series of items and others that required children to name individual items presented one at a time on a computer screen.
Several different models of the relationships among the three measurement domains were compared using confirmatory factor analysis to assess their fit to the data. The results supported the view that phonological awareness, memory, and naming are distinct yet correlated abilities. These results were replicated in our longitudinal sample of 216 children whose phonological abilities were assessed each year from kindergarten through fourth grade (Wagner et al., 1997). Prior to examining developmental relationships between phonological skills and reading, we performed tests of our measurement model for phonological abilities. The basic model proposed correlated factors for phonological awareness (both analysis and synthesis), memory, and rapid naming, and this model received strong support in the data. For example, the comparative fit index (an index that ranges from 0 for a terrible fit to 1.0 for a near-perfect fit) ranged from .97 to .99 for the models tested at different ages. When taken together, these studies provide strong support for conceptualizing phonological awareness, phonological memory, and rapid automatic naming tasks as measures of separate although correlated abilities.
The evidence that these phonological abilities are causally related to the growth of word reading skill is strongest for phonological awareness, next strongest for rapid automatic naming ability, and weakest for phonological coding in working memory. The evidence for phonological awareness’ role in reading development comes from both standard and causal modeling studies of longitudinal–correlational data (Mann, 1993; Stanovich, Cunning-ham, & Cramer, 1984; Wagner, Torgesen, & Rashotte, 1994; Wagner et al., 1997), from comparisons of reading-disabled and normal readers using reading-level-match designs (Bowey et al., 1992), and from true experiments that show an effect of training phonological awareness on subsequent reading ability (Lundberg, Frost, & Peterson, 1988; Torgesen, Morgan, & Davis, 1992). Two of these categories of evidence are available for rapid naming’s role in reading development: standard and causal modeling analyses of longitudinal–correlational data (Felton & Brown, 1990; Wagner et al., 1994, 1997; Wolf & Goodglass, 1986) and differences between normal and reading-disabled children using reading-level-match designs (Bowers et al., 1994).
Although individual differences in verbal short-term memory can be shown to predict subsequent reading development (Brady, 1991; Mann & Liberman, 1984), individual differences on these tasks do not appear to explain unique variance in reading growth beyond that explained by phonological awareness and rapid naming ability (Wagner et al., 1994, 1997). Furthermore, the evidence for special impairment of reading-disabled children on verbal short-term memory tasks from reading-level-match designs has been consistently negative (Pennington, Van Orden, Kirson, & Haith, 1991; Stanovich & Siegel, 1994). However, as Torgesen (1995) pointed out, most of the prominent case studies of phonological dyslexia in both children and adults report limitations in verbal short-term memory as one of the prominent cognitive characteristics of these subjects. Clearly, further research is required to determine what unique role, if any, problems in phonological coding in working memory play in causing reading disabilities. Although the three phonological abilities under discussion here appear to differ in their importance as unique causes of reading disabilities, because of their consistent relationships to one another, all of them are included in our discussion of stability. We turn now to an assessment of the consistency of individual differences in these abilities across development.
STABILITY OF PHONOLOGICAL ABILITIES DURING THE ELEMENTARY SCHOOL PERIOD
In all of our research on the development of reading-related phonological processes, we have employed multiple measures of each construct. This makes possible the use of latent variables in our analyses. A latent variable represents the common or shared variance among the several tests that are used to measure each construct. Because they represent only the variance common across tasks, latent variables provide relatively error-free measurement by excluding task-specific sources of error. When assessed as latent variables, reading-related phonological abilities show a high degree of stability across the elementary school period.
Stability of Phonological Abilities When Assessed as Latent Variables
Table 7.1 (Wagner et al., 1997) presents an index of stability for each variable (because analytic and synthetic phonological awareness were so highly correlated after second grade, they are combined into a single variable in these analyses) from kindergarten through fourth grade. Although analagous to correlation coefficients, these indexes are actually standardized maximum likelihood estimates of covariances among latent variables (Bollen, 1989). Their interpretation is very similar to standard correlation coefficients (they can vary from 0.0 to 1.0, depending on the strength of the relationship among constructs), and they show that individual differences in phonological skills are more stable than those for word-level reading from kindergarten and first grade to later grades, although they are roughly similar in stability to reading beginning about second grade. For example, the first- to fourth-grade stability co-efficients for phonological variables vary between .86 and .72, whereas that for word-level reading over the same period is .62. In contrast, the second- to fourth-grade stability indexes for phonological variables vary between .94 and .82, whereas that for word-level reading is .87.
Stability of Phonological Processing Abilities and Word-Level Reading Measured as Latent Variables
From: Wagner, R. K., Torgesen, J. K., Rashotte, C. A., Hecht, S. A., Barker, T. A., Burgess, S. R., Donahue, J., & Gar on, T. (1997). Changing causal relations between phonological processing abilities and word-level reading as children develop from beginning to fluent readers: A five-year longitudinal study. Developmental Psychology, 33, 468–479.
Stability of Phonological Skills When Assessed as Unit-Weighted Composites
Although relationships among latent variables provide the most accurate estimate of the true degree of relationship among constructs or their true stability over time, the use of latent measures is still relatively uncommon in developmental studies of reading and related processing skills. In order to provide an estimate of the stability of phonological skills and word reading ability when they are measured as observed variables, and also to provide a comparison of their stability with a measure of general verbal ability (which could not be treated as a latent variable because we only had one measure for it), we calculated stability co-efficients among composite measures of phonological awareness, rapid naming, memory, and word-level reading skill. All composite variables were formed as an average of the standard scores for each task used to construct the composite. The tasks combined to form the measure of phonological awareness were phoneme elision, segmenting, and phoneme blending. Measures of rapid digit naming and rapid letter naming were combined to estimate rapid automatic naming ability; and digit span and verbatim memory for sentences were combined as the measure of phonological memory. The Word Identification and Word Attack subtests of the Woodcock Reading Mastery Test–Revised (Woodcock, 1987) were combined to form the measure of word-level reading ability. General verbal ability was estimated from the Vocabulary subtest of the Stanford Binet Intelligence Scale (4th ed.; Thorndike, Hagen, & Sattler, 1986). In this analysis, data was available from a final testing of the longitudinal sample when the children were in fifth grade. Two hundred and one children remained in the sample for the final testing point. The correlations of each variable with itself across the span from kindergarten to fifth grade are presented in Table 7.2. The general pattern of results from this table indicates that, over the period from kindergarten through fifth grade, individual phonological abilities are at least as stable as general verbal ability estimated by a test of vocabulary knowledge.
The values reported in Table 7.2 are very similar to those reported by Scarborough (1995) for the 6-year interval between second and eighth grades. For example, the second- to eighth-grade correlation for phonological awareness in her study was .49, for rapid naming it was .51, and for phonological memory it was .66.
Stability of Phonological Processing Abilities and Vocabulary Knowledge as Observed Variables
Consistency of Significant Disabilities in Phonological Processing and Word Reading
Thus far, we have presented evidence that individual differences in phonological abilities are very stable across the elementary school period when the full range of abilities are included in the analyses. This suggests that children who are weak in phonological skills at the beginning of school will continue to be relatively weak in these abilities in fourth or fifth grade, which increases the likelihood that children with phonologically based reading disabilities will show relatively consistent reading difficulties across grade levels in elementary school. In order to provide a more direct assessment of the stability of significant weaknesses in phonological performance across the elementary school period, we identified children in first grade and third grades who performed below the 10th percentile on each skill, and then examined their placement within the distribution of that skill 2 years later (in third and fifth grades, respectively). In other words, we examined the stability of a diagnosis of significant weakness in phonological skill and word reading ability over 2-year increments during early and late elementary school. The same composite measures of each phonological skill and word reading that were used in the previous analysis were used here. Because of slight variations in the distribution of these skills within the sample, and because of occasional tied scores at the 10th percentile cut point, the number of children identified as significantly disabled (bottom 10%) varied slightly (from 20 to 22 children) across abilities.
Table 7.3 presents the percentage of the disabled group at first and third grades that fell at each decile within the bottom half of the distribution of that skill 2 years later. For example, of the original 21 children identified as significantly weak in performance on phonological awareness tasks in first grade, 51% continued to fall in the bottom 10th percentile in third grade. Although this suggests that about half of the children who were classified as significantly disabled in their performance on phonological awareness tasks at the beginning of first grade no longer fell in this category in third grade, 75% of these children fell below the 30th percentile in third grade, and only one child achieved above-average performance. The corresponding numbers (percent of first-grade disabled children falling below the 30th percentile in third grade) for rapid automatic naming and phonological memory were very similar, 95% and 75%, respectively. The overall stability of phonological disabilities from Grade 3 to Grade 5 was very similar to that for the earlier developmental period. The stability of extremely low performance on word reading tasks was similar to that for the phonological measures, with the period from Grade 3 to Grade 5 showing particularly extreme stability.
It should be pointed out here that children in our longitudinal sample were assessed at the beginning of the first, second, and third grades, and at the end of the fourth and fifth grades. Thus, the stability estimates for the early grades encompass a 2-year span starting before the beginning of formal reading instruction. In contrast, the stability estimates for Grades 3 and 5 encompass almost 3 full school years.
The overall pattern of substantial stability in the relative performance of children with phonological weaknesses over the elementary school period is consistent with other reports of stability of reading disabilities over the same period. Although we could not locate other studies that examined consistency of extremely poor performance on phonological processing tasks per se, there are a number of studies that have examined the stability of word reading disabilities in elementary school. For example, Juel (1988), in a longitudinal study of reading growth from first to fourth grades, reported that the probability of becoming an average reader by the end of fourth grade if a child fell within the bottom quartile in reading at the end of first grade was .13. In our sample, if we were to define average word reading ability as falling above the 40th percentile, then the probability of a child who is identified with severe word reading disabilities at the beginning of first grade performing in the average range at the beginning of third grade is .14. The same probability calculated for the period from the beginning of third grade to the end of fifth grade in our sample is 0.0. Thus, the conclusion from both these studies is that children who are significantly impaired in their phonological and word reading skills at one point in time during elementary school almost invariably are found to be poor or below average in these skills 2 years later. Furthermore, this stability is relatively consistent across both early and late periods of development.
Stability of Significant Disabilities in Phonological Processing and Word Reading Ability Over 2-Year Periods in Early and Late Elementary School
This conclusion about significant stability in the diagnosis of phonologically based reading disabilities is, however, inconsistent with the results of another widely cited study (Shaywitz, Escobar, Shaywitz, Fletcher, & Makuch, 1992) that reported very little stability in the diagnosis of specific developmental dyslexia over a similar period in elementary school. These investigators, for example, reported that only 7 of 25 (28%) students diagnosed as dyslexic in first grade received the same diagnosis in third grade, and only 14 of 30 (46%) diagnosed in third grade received the same diagnosis in fifth grade. These data were widely interpreted as suggesting that specific developmental dyslexia (or phonologically based reading disabilities) is an extremely unstable disability, so that there is a high probability a child might “have it” at one point and “not have it” at another point in development.
However, there are at least two reasons to discount this interpretation within the present context, in which our concern is with the stability of phonological processing and reading levels themselves, rather than the stability of a diagnosis that involves arbitrary cutpoints in the distibution of reading skills. First, what Shaywitz and her colleagues assessed was the stability of a discrepancy definition of reading disabilities that involved a difference between measured intelligence and reading level. Difference scores, such as those used to identify discrepancies, are statistically less reliable than the scores for absolute level of reading ability or level of intelligence taken by themselves. Second, the earlier study was conducted to examine classification reliability, which can be easily influenced by small amounts of movement across the arbitrary cutpoints that are used to determine whether a child receives the diagnosis or not. In our own data, if we analyze simple classification reliability and decide that a child must perform in the bottom 10th percentile in word reading ability to be classified as reading disabled, then our results become more similar to those of Shaywitz et al. (1992). For example, only 36% of children scoring in the bottom 10th percentile in first grade do so in third grade, and 75% of children with severe reading problems in third grade continue to show the same level of impairment in fifth grade. These stability estimates are slightly higher than those reported in the Shaywitz et al. (1992) study, most probably because they used a cutpoint for discrepancy score, whereas we used a cutpoint only for reading level. However, the important point about stability is the one made in previous paragraphs: Very few of the children identified with severe reading problems in first or third grade actually achieve average reading levels in subsequent years.
Factors Related to the Stability of Phonological Processes
We began this section by suggesting that evidence confirming the stability of phonological processing across a significant developmental span would provide additional support for the theory of phonologically based reading disabilities. The cause of an academic difficulty that is as difficult to eliminate as reading disabilities should be a relatively stable characteristic of the child. Our statement is only true, however, if it is combined with evidence that early individual differences in phonological processing are more strongly related to later growth in word reading ability than the other way around. In other words, because it is relatively widely accepted that learning to read stimulates the growth of some phonological skills (phonological awareness in particular; Brady & Shankweiler, 1991), it is possible that the stability of phonological processing skills is primarily the product of stability in the distribution of word reading ability in our sample. Although there may be some truth to this argument, it cannot be the whole story, because our analyses within this same sample (Wagner et al., 1994, 1997) have consistently shown more powerful longitudinal prediction of growth from early phonological skills to later reading skills than from individual differences in reading to subsequent growth in phonological ability. In other words, our analyses of causal relationships among these variables suggest that causal influences are stronger from phonological processing to reading than from reading to phonological processing.
It also must be pointed out that the stability estimates we have reported may indicate as much about the educational environment of the children in our study as they do about the nature of phonological processes themselves. Specifically, it is possible that our data overestimate the general degree of stability in phonological abilities in young children because of a relatively weak program of instruction in word reading skills in the school district from which this sample was taken. For example, at the end of fifth grade (approximately 5.7 grade level), the average grade level score on the Word Attack subtest of the Woodcock Reading Mastery Test–Revised (Woodcock, 1987) was 4.1, and 40% of the sample achieved scores on this test that were two or more grade levels below their current grade placement. The Word Attack subtest is a direct measure of children’s ability to apply “phonics” knowledge (letter-sound correspondences, phoneme blending skills) to reading novel words. This level of performance for our sample indicates either that our children as a whole had greater difficulty learning phonological skills in reading, or that they were less consistently supported in their development of these skills, than the national sample from which the test’s norms were calculated. Although it is difficult to predict the precise effects that more consistent instruction and general support for the development of word-level reading skills would have on the stability of individual differences in phonological abilities, it seems likely that more effective instruction for low-ability children might reduce these estimates slightly if it were able to accelerate the development of skills in this area. We turn now to an examination of instructional effects for phonological awareness, which indicate that this core phonological skill can be significantly modified by direct instruction, but also that there are significant individual differences in response to this training.
Response to Training in Phonological Awareness
With the growing consensus about the importance of individual differences in phonological skills as causes of variation in response to early reading instruction, there has arisen the hope that early training in these skills might significantly reduce the incidence of reading disabilities in young children (Berninger, Thalberg, DeBruyn, & Smith, 1987). This hope has been supported by findings that phonological awareness, in particular, can be substantially improved in groups of kindgarten and first-grade children through direct instruction using oral language activities. For example, a recent meta-analysis of 13 phonological awareness training studies (Wagner, Torgesen, & Rashotte, 1993) found an average effect size of 1.23 standard deviation units on measures of phonological awareness after an average of only 9 hours of training. A recent study (Foster, Erickson, Foster, Brinkman, & Torgesen, 1994) using computerized activities to stimulate phonological awareness reported an effect size of 1.05 standard deviation units after only 4½ total hours of training! A number of studies (Ball & Blachman, 1991; Barker & Torgesen, 1995; Cunningham, 1990; Lundberg et al., 1988) have also demonstrated significant average effects on children’s subsequent reading development following training in phonological awareness. Taken together, the data on ease of training in phonological awareness associated with subsequent improvements in reading growth provide support for hopes that early training in phonological skills may have a serious impact on the reading problems of many children.
However, a crucial question that is not consistently addressed in the training literature on phonological awareness is whether this training simply “hothouses,” or accelerates, the growth of children who would be normal readers anyway, or whether it actually reduces the risk of reading failure for children with the weakest phonological abilities before training. For example, in commenting informally on his large-scale and generally successful training study using random samples of kindergarten children (Lundberg et al., 1988), Lundberg (1988) indicated that a substantial number of lower-ability children did not appear to significantly profit from the training. In our own work to develop a phonological awareness training program for kindergarten children (Torgesen et al., 1992), we found that approximately one third of a group of low-ability children did not respond (by showing meaningful growth in outcome measures) to a training program in which groups of three to four children received approximately 7 hours of training spread over 7 to 8 weeks.
Intrigued by this significant range of individual differences in response to training in phonological awareness, we conducted a study that was specifically designed to examine the child characteristics that predict growth in response to training (Torgesen & Davis, 1997). Sixty kindergarten children from two elementary schools with historically low achievement in reading were selected because of relatively low performance on a screening measure for phonological awareness. They were then provided with 16 hours of training that was spread over 12 weeks. The children were trained in small groups (three or four children), and the training focused on both analytic (recognizing and thinking about individual sounds in words) and synthetic (blending individual sounds into words) awareness skills. Children were taught to blend phonemes together to make words, to identify the positions of specific phonemes in words, and to pronounce individual phonemes in words presented as wholes. These activities were supported by the use of pictures to represent words, colored blocks to provide concrete representations of phonemes in words, and a variety of gamelike activities to maintain interest. During approximately the last 3 weeks of training, the children were taught the letters that represented a small group of the phonemes they had practiced with during their oral language training. They were then shown how to use their letter knowledge, along with their segmenting and blending skills, to read a small group of real words.
For the entire group, the effect size (compared to a no-treatment control group) for growth in analytic phonological awareness (as measured by ability to separately pronounce the individual sounds in words) was 1.35, whereas that for synthetic awareness (ability to blend separately presented phonemes into words) was 1.85. These are substantial overall training effects, and they are very consistent with average effects that are typically observed in training studies. However, 21 of the 60 children in the training group scored only 1 or 0 on the segmenting posttest; they had gained no reliable analytic skills as a result of the training. After determining that there was a significant overall training effect for both analytic and synthetic skills, our next step was to examine the individual child characteristics that best predicted response to training.
This analysis used hierarchical linear modeling to estimate a growth curve for each child in both analytic and synthetic phonological awareness. We were able to estimate only the linear component of growth, because we had available only three measures of performance: at the beginning, middle, and end of training. The slopes of these growth curves were predicted by variables from four categories: pretest levels of phonological awareness; other phonological processing skills—rapid automatic naming for digits and phonological short-term memory; beginning reading/spelling skills; and general verbal ability.
Consistent with our earlier work showing that analytic and synthetic skills are separate but correlated abilities in young children (Wagner, Torgesen, Laughon, Simmons, & Rashotte, 1993), different variables predicted growth in these two types of phonological awareness. Pretest measures of rapid automatic naming and invented spelling (a measure of both analytic awareness and letter-sound knowledge) were the best predictors of growth in blending skill, whereas measures of invented spelling and general verbal ability were the strongest predictors of growth in analytic ability.
These results indicate that children in our sample who had relative weaknesses in rapid automatic naming ability, phonological awareness, and knowledge of letter-sound correspondences were those who profited least from the training in phonological awareness. Because all of these characteristics have been associated with phonologically based reading disabilities, it seems likely that a large proportion of reading-disabled children would fall into the group of children who did not profit from the otherwise effective training provided in this study. This suggests that, if training in phonological awareness is to be successful in reducing the number of children who develop reading disabilities, it may require training procedures that go beyond those typically found in the research literature, in terms of both their explicitness and intensity. We turn now to a consideration of the practical implications, for both identification and treatment, of the research findings we have considered thus far.
Practical Applications of Research Knowledge to Diagnoses and Intervention for Reading-Disabled Children
The information presented thus far on the stability of phonological skills during the elementary school years, as well as indications that disabilities in these areas may not be easily overcome with moderate-level interventions, has implications for practice in two important areas. First, the generally strong relationships between early individual differences in phonological processing skill and later individual differences in both phonological skills and reading suggests the possibility for advances in our ability to identify children at risk for specific reading disabilities prior to the start of reading instruction. Rather than waiting for intervention to begin once children experience reading failure in first or second grade, we now have the potential to develop theoretically and practically meaningful instruments to identify children for placement in preventive instructional programs. Second, information about the stability of phonological skills in large random samples of children, as well as resistance to treatment of phonological disabilities in low-ability children, suggest the need for a re-examination of intervention procedures for children with phonologically based reading disabilities.
Use of Phonological Measures for Early Identification
Because studies of the relationships between phonological skills and word reading growth generally indicate that phonological awareness is more strongly related to early reading growth than other types of phonological skill (Fletcher et al., 1994; Wagner et al., 1994), there have been a number of recent attempts to use early measures of phonological awareness to identify children who will later fail in reading (Catts, 1996). These studies have measured phonological awareness either during kindergarten, or at the beginning of first grade, and report overall identification accuracy’s ranging from 75.6% to 99%. With the exception of two studies that measured phonological skills at the beginning of first grade, however, these studies have universally had high numbers of false positives (children who are predicted to be poor readers, but turn out to be good readers), with rates ranging from 23% to 69%. Scarborough (1996) concluded that, in order to more accurately identify specific children as at risk of reading failure, it may be useful to supplement measures of phonological awareness with measures of other phonological processes and emergent reading skills.
Using data from the longitudinal study reported on earlier in this chapter, we (Torgesen & Wagner, 1995) used a combination of three measures taken in the first semester of kindergarten to identify children at risk of being in the bottom 10% of children in word reading ability during the first semester of second grade. The predictive measures we used were letter-name knowledge (how many of the 26 uppercase letters could be named), a measure of phonological awareness (phoneme elision—“Say mat without saying the /m/ sound.”), and rapid automatic naming ability for digits (six digits were randomly arranged in six rows for a total of 36 digits). These measures were combined together using logistic regression procedures, and 23 children were identified (of the 240 children in the sample) who had a probability greater than 50% of being in the bottom 10% in second grade. Of these 23 children, 14 actually obtained word reading scores in the bottom 10% in second grade. Of the nine “false positives” in this group, five were in the bottom 20%, two in the bottom 30%, and two were slightly above-average readers. The overall accuracy rate in this analysis was 92%, and the rate of false positives was only 4%. However, the predictive procedure we used failed to identify 10 of the 24 children who ended up in the bottom 10% of readers in second grade, which gave us a 42% false negative rate. Although this balance between false positives and false negatives was appropriate for our purpose, which was to select children for a study of instructional procedures to prevent reading disabilities, the rate of false negatives would be unacceptably high in practice. We are currently working on a different predictive equation, using a few more indicators, that will examine more completely the potential of these kinds of measures to accurately identify specific children who are most in need of supplemental instruction to prevent the early emergence of serious reading difficulties.
One issue that currently limits research on the early identification of children at risk of phonologically based reading disabilities is the lack of standardized measures in this area. There are a number of standardized (in terms of administration procedures) instruments available to measure phonological awareness, such as the Lindamood Auditory Conceptualization Test (Lindamood & Lindamood, 1979), the Test of Phonological Awareness (Torgesen & Bryant, 1994), the Test of Awareness of Language Segments (Sawyer, 1987), the Phonological Awareness Profile (Robertson & Salter, 1995), and the Yopp-Singer Test of Phoneme Segmentation (Yopp, 1995). However, only one of these tests includes more than one measure of phonological awareness, and none of them allow simultaneous assessment of phonological skills other than awareness. This latter constraint limits the kind of multivariate procedures recommended by Scarborough (1996) that are most likely to obtain the high identification rates that will make early screening and intervention for phonologically based reading disabilities practically feasible.
In our research program, we are currently in the processes of developing the Comprehensive Test of Phonological Processes in Reading (CTPPR), which will provide wide range assessment of all three areas of phonological skill discussed in this chapter. The CTPPR will contain multiple subtests to assess phonological awareness, phonological memory, and rapid automatic naming skill. We have carried out preliminary studies of the items on these subtests as well as its concurrent validity, and the test will be submitted to a national standardization during the 1996–1997 school year.
Another issue involving the use of phonological processing measures to identify children at risk for reading failure is that they are likely to identify a somewhat different group than is currently being served in programs for children with reading disabilities. We are suggesting the idea that children might be identified for special preventive instruction in reading on the basis of low performance on measures that assess a set of cognitive markers (perhaps supplemented by demographic markers) for reading disability. In contrast, procedures currently in use, and supported by a long tradition in the field of learning disabilities, identify children for special educational help in reading on the basis of a discrepancy between their general ability (as measured by a full-scale IQ test) and their reading achievement. This practice arose, in part, because of assumptions about differences between children whose reading level was discrepant versus nondiscrepant from their general intelligence, and also from a lack of knowledge about the critical cognitive factors responsible for reading disabilities in children (Torgesen, 1993). Recent research has established that the basic causes of word reading difficulties in readers of low general intelligence are indistinguishable from those responsible for reading problems in children of average and high general intelligence (Fletcher, Francis, Rourke, Shaywitz, & Shaywitz, 1992; Fletcher et al., 1994; Pennington, Gilger, Olson, & DeFries, 1992; Stanovich & Siegel, 1994). Thus, it seems appropriate that children with similar problems in acquiring word reading ability should have available similar interventions early in their schooling.
However, it is true that children whose word reading skills are discrepant versus nondiscrepant from their general intelligence will differ on cognitive skills that lie outside the word recognition module (Ellis & Large, 1987; Stanovich & Siegel, 1994). Children whose general ability is significantly higher than their ability to acquire word reading skills are likely to have higher reading comprehension levels if the basic deficit in word recognition skill can be overcome. This argument implies that children of high and low general intelligence do have a different prognosis with regard to the ultimate level of reading skill they are likely to attain, given equivalent instruction. It also implies that children of low general ability (specifically broad verbal ability) may require additional intensive instruction to acquire the general verbal knowledge and skill that supports good reading comprehension. Because of the sheer amount of knowledge and skill to be acquired, this latter deficiency is likely to be much more difficult to overcome than are problems in word reading.
The Impact of Intensive Interventions
We are currently engaged in a 5-year project to study interventions that are focused on the word reading difficulties of children with phonologically based reading disabilities. It is clear both from studies of long-term outcomes (Horn, O’Donnell, & Vitulano, 1983) for students with reading disabilities and from studies of short-term effectiveness of special education programs (Kaufman, 1993) that public education is not currently meeting the educational needs of these children. In our own longitudinal study, for example, none of the children with word reading skills below the 10th percentile in third grade were reading above the 30th percentile in the fifth grade.
There are, however, scattered clinical reports indicating that appropriate instruction, if offered with sufficient intensity, can frequently produce dramatic improvements in the fundamental reading skills of children who were previously making little progress in learning to read (Clark & Uhry, 1995). These clinical reports have been recently supplemented with research evidence that it is possible to substantially improve the word reading skills of children with specific reading disabilities by using instructional programs that explicitly teach the structure of language at the word level coupled with extensive review and practice (Alexander, Anderson, Heilman, Voeller, & Torgesen, 1991; Lovett, Borden, Lea Lacerenza, Benson, & Brackstone, 1994; Olson, Wise, Johnson, & Ring, 1997).
These reports are consistent with data we have obtained from our current study of remedial instruction with reading-disabled children, in which we have found quite dramatic short-term effects on children’s word reading abilities as a result of two different types of intensive intervention. This study focused on children who were already identified by their school district as learning disabled in reading, and, from this group, we selected children with the lowest scores on word-level reading skills and phonological processing measures. These selected children received 80 hours of one-on-one tutorial instruction using either of two instructional methods. One method, called the Auditory Discrimination in Depth (ADD; Lindamood & Lindamood, 1984) program, provides very explicit instruction in phonological awareness by helping children discover the articulatory gestures that are associated with each phoneme in the English language. This oral awareness instruction is followed by extensive practice in monitoring the number, identity, and sequence of sounds in words by using both articulatory and auditory information. The program then teaches children to apply what they have learned in the phonological awareness activities to reading and spelling words. Children received extensive decontextualized practice in reading both real and nonwords, and they were engaged in supported reading experiences with text that were phonetically controlled in the beginning.
The other instructional method is called embedded phonics (EP). This program provided instruction in reading that was oriented more toward the whole word and reading for meaning than the ADD program. Children spent the majority of their time reading either basal or trade books, and unusual reading vocabulary in these books was directly pretaught. Phonological awareness was stimulated during the writing of some of these words, in which the teacher encouraged the child to think about the number of sounds in the word before beginning to write. Phonics skills were directly taught, but they were not taught sequentially or drilled using decontextualized activities. During the introduction of new sight words, the phonetic patterns used in the words were introduced. Children were also explicitly taught to utilize both grapheme-based and context-based cues for the identity of words in text.
Both of these methods contained direct instruction in phonics, and they both included activities designed to stimulate phonological awareness. The fundamental contrast between them was in the degree of explicitness and amount of total instructional time focused on building phonological awareness and practicing alphabetic decoding skills.
We now have pre- and posttest outcome data for 58 students: 29 in the ADD program, and 29 who were randomly assigned to the EP method. The average age of these children when they began our program was 121 months, and their average full-scale IQ was 98, with a range from 82 to 128. There are 40 males and 18 females in the sample, with the racial balance being predominantly White. About 55% of the children have received a comorbid diagnosis of attention deficit disorder, and these children have all received stimulant medication during the intervention.
The children were assessed with a variety of reading and language measures, and their continued development will be assessed for 2 years following the end of our intervention. Measurement of short-term changes in their word reading ability using a standardized measure (Woodcock Reading Mastery Test–Revised; Woodcock, 1987) indicates substantial improvement for children receiving instruction from both methods. For example, children in the ADD group showed an average gain in the ability to apply phonics skills to reading novel words (Word Attack subtest) of 27 standard score points, with the corresponding gain for the EP group being 20 points. These gains in performance correspond to movement from the 2nd percentile to the 37th and 25th percentiles, respectively. The ADD group showed significantly more improvement on this measure than the EP group. For ability to read increasingly difficult real words (Word Identification subtest), average improvement for both groups was approximately 14 standard score points, which corresponds to movement from the 2nd to the 10th percentile. In the area of reading comprehension, the children finished at about the 27th percentile. In order to illustrate graphically the significant changes in these children’s word-level reading abilities within a short time period, their average raw scores on the Word Attack and Word Identification subtests are plotted in Fig. 7.1 as a function of their age in months. The top dotted line plots data from the national standardization for the test, to indicate normal growth on these measures. These graphs assume a common starting point for all children at the beginning of first grade, and they represent pre-, mid-, and posttest data on the intervention sample during the 8 weeks of their intervention programs. We currently have 1 year follow-up data on 34 children, and this data is also shown in Fig. 7.1.
FIG. 7.1. Change in raw scores on Word Attack and Word Identification subtests as a result of intensive instructional interventions.
Of course, the ultimate effectiveness of these interventions is dependent on more than short-term gains in decontextualized word reading ability. One important test is whether the children have acquired generalizable skills that will assist them to continue to grow in their reading ability after the intervention is concluded. Our preliminary answer to that question for these interventions is that the large majority (30 of 34 children for whom 1-year follow-up data is presently available) of children do continue to show improvements (measured by gains in standard scores) in Word Attack and Word Identification performance 1 year later. Equally important to the question of continued growth in word reading ability is whether growth in these skills leads to significant and continuing gains in reading comprehension and reading fluency. Although we are collecting data on this issue, they are not yet available for report. A recent review of intervention research in reading by Olson et al. (1997) suggests that the issue of growth in fluency and strong generalization of remediated word reading skills to text processing must be examined carefully in future research.
CONCLUDING COMMENTS
The major point of this chapter is that, under typical instructional conditions in elementary school, individual differences in phonological processing and word reading ability are extremely stable over the period from first through fifth grades. This stability has two important implications for professionals concerned with assisting the growth of all children in reading. First, it should be possible to utilize phonological processing measures, in combination with other measures of emergent reading skills, to accurately identify children who will require additional instructional support to avoid early failure in learning to read. Although the question is still open for additional research, it seems likely at this point that these children will require instruction that is more explicit, more intensive, and more supportive than is currently available in most school settings (Torgesen, 1996). This second implication derives from the fact that children with the lowest abilities in the phonological domain do not easily respond to conventional levels of intensity and explicitness of instruction in the skills required to accurately identify words in text. However, several recent research reports do indicate that a number of different approaches can be effective in significantly enhancing word-level reading skills in children with phonologically based reading disabilities if they are offered with sufficient intensity. Our immediate next goals should be to explore ways of delivering the “special education” these children need within the school setting along with continued development of this instruction, so that it leads to broadly functional reading skills.
ACKNOWLEDGMENTS
The research reported in this manuscript was supported by grant numbers HD23340 and HD30988 from the National Institute of Child Health and Human Development, and by grants from the National Center for Learning Disabilities and the Donald D. Hammill Foundation.
REFERENCES
Alexander, A., Anderson, H., Heilman, P. C., Voeller, K. S., & Torgesen, J. K. (1991). Phonological awareness training and remediation of analytic decoding deficits in a group of severe dyslexics. Annals of Dyslexia, 41, 193–201.
Baddeley, A. D. (1991). Human memory: Theory and practice. London: Allyn & Bacon.
Ball, E. W., & Blachman, B. A. (1991). Does phoneme awareness training in kindergarten make a difference in early word recognition and developmental spelling? Reading Research Quarterly, 26, 49–66.
Barker, T. A., & Torgesen, J. K. (1995). An evaluation of computer-assisted instruction in phonological awareness with below average readers. Journal of Educational Computing Research, 13, 89–103.
Berninger, V., Thalberg, T., DeBruyn, I., & Smith, R. (1987). Preventing reading disabilities by assessing and remediating phonemic skills. School Psychology Review, 16, 554–565.
Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley.
Bowers, P., Golden, J., Kennedy, A., & Young, A. (1994). Limits upon orthographic knowledge due to processes indexed by naming speed. In V. W. Berninger (Ed.) The varieties of orthographic knowledge I: Theoretical and developmental issues (pp. 173–218). Dordrecht, The Netherlands: Kluwer.
Bowers, P. G., & Swanson, L. B. (1991). Naming speed deficits in reading disability: Multiple measures of a singular process. Journal of Experimental Child Psychology, 51, 195–219.
Bowers, P. G., & Wolf, M. (1993). Theoretical links betwen naming speed, precise timing mechanisms and orthographic skill in dyslexia. Reading and Writing: An Interdisciplinary Journal, 5, 69–85.
Bowey, J. A., Cain, M. T., & Ryan, S. M. (1992). A reading-level design study of phonological skills underlying fourth-grade children’s word reading difficulties. Child Development, 63, 999–1011
Brady, S. A. (1991). The role of working memory in reading disability. In S. A. Brady & D. P. Shankweiler (Eds.), Phonological processes in literacy (pp. 129–152). Hillsdale, NJ: Lawrence Erlbaum Associates.
Brady, S., Poggie, E., & Rapala, M. M. (1989). Speech repetition abilities in children who differ in reading skill. Language and Speech, 32, 109–122.
Brady, S., & Shankweiler, D. (1991). Phonological processes in literacy: A tribute to Isabelle Y. Liberman. Hillsdale, NJ: Lawrence Erlbaum Associates.
Bruck, M. (1992). Persistence of dyslexics’ phonological awareness deficits. Developmental Psychology, 28, 874–886.
Catts, H. (1996, March). Phonological awareness: A key to detection. Paper presented at conference titled The Spectrum of Developmental Disabilities XVIII: Dyslexia, at the Johns Hopkins Medical Institutions, Baltimore, MD.
Clark, D. B., & Uhry, J. K. (1995). Dyslexia: Theory and practice of remedial instruction (2nd ed.). Baltimore: York.
Cunningham, A. E. (1990). Explicit versus implicit instruction in phonemic awareness. Journal of Experimental Child Psychology, 50, 429–444.
Denckla, M. B., & Rudel, R. (1976). Naming of object drawing by dyslexic and other learning disabled children. Brain and Language, 3, 1–16.
Ellis, N., & Large, B. (1987). The development of reading: As you seek so shall you find. British Journal of Psychology, 78, 1–28.
Felton, R., & Brown, I. S. (1990). Phonological processes as predictors of specific reading skills in children at risk for reading failure. Reading and Writing: An Interdisciplinary Journal, 2, 39–59.
Fletcher, J. M., Francis, D. J., Rourke, B. P., Shaywitz, S. E., & Shaywitz, B. A. (1992). The validity of discrepancy-based definitions of reading disabilities. Journal of Learning Disabilities, 25, 555–561.
Fletcher, J. M., Shaywitz, S. E., Shankweiler, D. P., Katz, L., Liberman, I. Y., Stuebing, K. K., Francis, D. J., Fowler, A. E., & Shaywitz, B. A. (1994). Cognitive profiles of reading disability: Comparisons of discrepancy and low achievement definitions. Journal of Educational Psychology, 86, 6–23.
Foster, K. C., Erickson, G. C., Foster, D. F., Brinkman, D., & Torgesen, J. K. (1994). Computer administered instruction in phonological awareness: Evaluation of the DaisyQuest program. Journal of Research and Development in Education, 27, 126–137.
Gathercole, S. E., & Baddeley, A. D. (1993). Phonological working memory: A critical building block for reading development and vocabulary acquisition? European Journal of Psychology of Education, 8, 259–272.
Horn, W. F., O’Donnell, J. P., & Vitulano, L. A. (1983). Long-term follow-up studies of learning disabled persons. Journal of Learning Disabilities, 16, 542–555.
Hulme, C. (1988). Short-term memory development and learning to read. In M. Gruneberg, P. Morris, & R. Sykes (Eds.), Practical aspects of memory: Current research and issues. Vol. 2. Clinical and educational implications (pp. 234–271). Chichester, England: Wiley.
Jorm, A. F. (1983). Specific reading retardation and working memory: A review. British Journal of Psychology, 74, 311–342.
Juel, C. (1988). Learning to read and write: A longitudinal study of 54 children from first through fourth grades. Journal of Educational Psychology, 80, 437–447.
Kaufman, J. M. (1993). How we might achieve the radical reform of special education. Exceptional Children, 60, 6–16.
Liberman, I. Y., Shankweiler, D., & Liberman, A. M. (1989). The alphabetic principle and learning to read. In D. Shankweiler & I. Y. Liberman (Eds.), Phonology and reading disability: Solving the reading puzzle (pp. 1–33). Ann Arbor: University of Michigan Press.
Lindamood, C. H., & Lindamood, P. C. (1979). Lindamood auditory conceptualization test. Austin, TX: PRO-ED.
Lindamood, C. H., & Lindamood, P. C. (1984). Auditory discrimination in depth. Austin, TX: PRO-ED.
Lovette, M. W., Borden, S. L., Lea Lacerenza, T. D., Benson, N. J., & Brackstone, D. (1994). Treating the core deficits of developmental dyslexia: Evidence of transfer of learning after phonologically- and strategy-based reading training programs. Journal of Educational Psychology, 30, 805–822.
Lundberg, I. (1988). Preschool prevention of reading failure: Does training in phonological awareness work? In R. L. Masland & M. W. Masland (Eds.), Prevention of reading failure (pp. 163–176). Parkton, MD: York.
Lundberg, I., Frost, J., & Peterson, O. (1988). Effects of an extensive program for stimulating phonological awareness in pre-school children. Reading Research Quarterly, 23, 263–284.
Mann, V. A. (1993). Phoneme awareness and future reading ability. Journal of Learning Disabilities, 26, 259–269.
Mann, V. A., & Liberman, I. Y. (1984). Phonological awareness and verbal short-term memory. Journal of Learning Disabilities, 10, 592–599.
Olson, R. K., Wise, B., Johnson, M., & Ring, J. (1997). The etiology and remediation of phonologically based word recognition and spelling disabilities: Are phonological deficits the “hole” story? In B. Blachman (Ed.), Foundations of reading acquisition and dyslexia (pp. 305–326). Mahwah, NJ: Lawrence Erlbaum Associates.
Pennington, B. F., Gilger, J. W., Olson, R. K., & DeFries, J. C. (1992). The external validity of age-versus IQ-discrepancy definitions of reading disability: Lessons from a twin study. Journal of Learning Disabilities, 25, 562–573.
Pennington, B. F., Van Orden, G., Kirson, D., & Haith, M. (1991). What is the causal relation between verbal STM problems and dyslexia? In S. A. Brady & D. P. Shankweiler (Eds.), Phonological processes in literacy: A tribute to Isabelle Y. Liberman (pp. 173–186). Hillsdale, NJ: Lawrence Erlbaum Associates.
Rack, J. P., Snowling, M. J., & Olson, R. K. (1992). The nonword reading deficit in developmental dyslexia: A review. Reading Research Quarterly, 27, 29–53.
Robertson, C., & Salter, W. (1995). Phonological awareness profile. East Moline, IL: LinguiSystems.
Rudel, R. G., Denckla, M. B., & Broman, M. (1978). Rapid silent response to repeated target symbols by dyslexic and nondyslexic children. Brain and Language, 6, 52–62.
Sawyer, D. J. (1987). Test of awareness of language segments. Austin, TX: PRO-ED.
Scarborough, H. S. (1995, April). Long-term prediction of reading skills: Grade 2 to grade 8. Paper presented at annual meetings of the Society for Research in Child Development, Indianapolis, IN.
Scarborough, H. S. (1996, March). Phonological awareness: Just a piece of the puzzle. Paper presented at conference titled The Spectrum of Developmental Disabilities XVIII: Dyslexia, at the Johns Hopkins Medical Institutions, Baltimore, MD.
Share, D. L., & Stanovich, K. E. (1995). Cognitive processes in early reading development: A model of acquisition and individual differences. Issues in Education: Contributions From Educational Psychology, 1, 1–57.
Shaywitz, S. E., Escobar, M. D., Shaywitz, B. A., Fletcher, J. M., & Makuch, R. (1992). Evidence that dyslexia may represent the lower tail of a normal distribution of reading ability. The New England Journal of Medicine, 326, 145–150.
Stanovich, K. E. (1988). Explaining the differences between the dyslexic and the garden-variety poor reader: The phonological-core variable-difference model. Journal of Learning Disabilities, 21, 590–604.
Stanovich, K. E., Cunningham, A. E., & Cramer, B. B. (1984). Assessing phonological awareness in kindergarten children: Issues of task comparability. Journal of Experimental Child Psychology, 38, 175–190.
Stanovich, K. E., & Siegel, L. S. (1994). The phenotypic performance profile of reading-disabled children: A regression-based test of the phonological-core variable-difference model. Journal of Educational Psychology, 86, 24–53.
Thorndike, R. L., Hagen, E. P., & Sattler, J. M. (1986). Guide for administering and scoring, the Stanford-Binet Intelligence Scale: Fourth edition. Chicago: Riverside.
Torgesen, J. K. (1993). Variations on theory in learning disabilities. In R. Lyon, D. Gray, N. Krasnegor, & J. Kavenagh (Eds.), Better understanding learning disabilities: Perspectives on classification, identification, and assessment and their implications for education and policy (pp. 153–170). Baltimore: Brookes.
Torgesen, J. K. (1995). A model of memory from an information processing perspective: The special case of phonological memory. In G. Reid Lyon (Ed.), Attention, memory, and executive function: Issues in conceptualization and measurement (pp. 157–184). Baltimore: Brookes.
Torgesen, J. K. (1996). Thoughts about intervention research in learning disabilities. Learning Disabilities: An Interdisciplinary Journal, 7, 55–58.
Torgesen, J. K., & Bryant, B. (1994). Test of phonological awareness. Austin, TX: PRO-ED.
Torgesen, J. K., & Davis, C. (1997). Individual difference variables that predict response to training in phonological awareness. Journal of Experimental Child Psychology, 63, 1–21.
Torgesen, J. K., Morgan, S. T., & Davis, C. (1992). Effects of two types of phonological awareness training on word learning in kindergarten children. Journal of Educational Psychology, 84, 364–370.
Torgesen, J. K., Rashotte, C. A., Greenstein, J., Houck, G., & Portes, P. (1987). Academic difficulties of learning disabled children who perform poorly on memory span tasks. In H. L. Swanson (Ed.), Memory and learning disabilities: Advances in learning and behavioral disabilities (pp. 305–334). Greenwich, CN: JAI.
Torgesen, J. K., & Wagner, R. K. (1995, May). Alternative diagnostic approaches for specific developmental reading disabilities. Manuscript prepared for the National Research Council’s Board on Testing and Assessment. Presented at a Workshop on IQ Testing and Educational Decision Making, Washington, DC.
Torgesen, J. K., Wagner, R. K., Rashotte, C. A., Burgess, S., & Hecht, S. (1997). The contributions of phonological awareness and rapid automatic naming ability to the growth of word reading skills in second to fifth grade students. Scientific Studies of Reading, 1, 161–185.
Wagner, R. K., Torgesen, J. K., & Rashotte, C. (1993, April). The efficacy of phonological awareness training for early reading achievement: A meta-analysis. Symposium presented at annual meetings of the American Educational Research Association, Atlanta, GA.
Wagner, R. K., Torgesen, J. K., Laughon, P., Simmons, K., & Rashotte, C. A. (1993). The development of young readers’ phonological processing abilities. Journal of Educational Psychology, 85, 1–20.
Wagner, R. K., Torgesen, J. K., & Rashotte, C. A. (1994). The development of reading-related phonological processing abilities: New evidence of bi-directional causality from a latent variable longitudinal study. Developmental Psychology, 30, 73–87.
Wagner, R. K., Torgesen, J. K., Rashotte, C. A., Hecht, S. A., Barker, T. A., Burgess, S. R., Donahue, J., & Garon, T. (1977). Changing causal relations between phonological processing abilities and word-level reading as children develop from beginning to fluent readers: A five-year longitudinal study. Developmental Psychology, 33, 468–479.
Wolf, M. (1991). Naming speed and reading: The contribution of the cognitive neurosciences. Reading Research Quarterly, 26, 123–141.
Wolf, M., & Goodglass, A. (1986). Dyslexia, dysnomia, and lexical retrieval: A longitudinal investigation. Brain and Language, 28, 154–168.
Woodcock, R. W. (1987). Woodcock reading mastery tests-revised. Circle Pines, MN: American Guidance Service.
Yopp, H. K. (1995). A test for assessing phonemic awareness in young children. The Reading Teacher, 49, 20–29.