Skip to main content
Elementary boy in yellow plaid shirt taking a test

Assessment and Evaluation

Research-Supported Assessment: Intervention Links for Reading and Writing

This article discusses current research-supported instructional practices in reading and writing. It also reviews alternatives to ability-achievement discrepancy in identifying students for special education services, as well as introduces the idea that ability-achievement discrepancies should be based on specific cognitive factors that are relevant to specific kinds of learning disabilities rather than Full Scale IQ.

On this page:

Visionaries in the field of school psychology have written eloquently over the years (e.g., Reschly & Wilson, 1995; Shapiro, 1996) about expanding the role of the school psychologist beyond that of assessor of ability-achievement discrepancies to that of the intervention specialist concerned with prevention and treatment, as well as assessment, of learning disabilities. That vision is likely to become a reality with the forthcoming reauthorization of IDEA, which most likely will not rely on IQ-achievement discrepancy as a means of identification for learning disabilities services because of the many problems research has found with this definition of learning disability (Lyon et al., 2001). 

The reauthorization of IDEA will also probably allow alternative ways to identify students for special education and emphasize early intervention and prevention of learning problems. Recent federal initiatives related to reading such as Reading First and No Child Left Behind are emphasizing the importance of using research-supported reading methods in early intervention (see Foorman, 2003).

In this article, we first review alternatives to ability-achievement discrepancy in identifying students for special education services. These alternatives supplement rather than replace ability-achievement discrepancy. We introduce the idea that ability-achievement discrepancies should be based on specific cognitive factors that are relevant to specific kinds of learning disabilities rather than Full Scale IQ. Then we discuss current research-supported instructional practices in reading and writing.

Finally, we describe the three-tier model developed at the University of Washington as one way to conceptualize school psychology service delivery to implement system-wide change in meeting the needs of all students in general and special education. We explain what research has found about effective early intervention in reading and writing (tier 1), preparation of students to meet standards on high-stakes tests (tier 2), and treatment of biologically-based, persisting severe reading and writing disorders (tier 3). We relate this model to another three-tier model developed at the University of Houston and the University of Texas at Austin for alternative delivery of general education and special education services.

Alternatives to IQ-achievement discrepancy

Researchers, school practitioners, professors in professional preparation programs, parents and other constituencies participated in a national summit on learning disabilities in 2001 (for proceedings, see Bradley, Danielson, & Hallahan, 2002). This summit confirmed that ability-achievement discrepancy

  • Has been used as the hallmark of learning disability
  • Has increased by 300% the number of children qualified for special education services since 1976 under the learning disability category
  • Contributes to the disproportionate minority representation in special education (Reschly, 2002; Reschly, 1997)
  • Generally lacks research support as the only way to identify all students needing individually tailored instruction

School psychologists are asking if not ability-achievement discrepancy, then what? Four possibilities to consider are discussed next.

Process assessment

The early research that discredited assessment and treatment of processes underlying learning disabilities was not focused on the processes that recent research has shown are relevant to reading and writing (Berninger, 1998a). The issue of how best to apply assessment of processes related to reading in identification of students with reading problems (Torgesen, 1979) has not been fully resolved (Torgesen, 2002). Yet, researchers have made considerable progress in identifying which processes are most relevant to

  • Learning to read and spell words
  • Predicting who will have reading and spelling problems if not provided early intervention
  • Improving reading or writing by training processes related to these literacy skills

Relevant processes in reading, writing and spelling

The processes relevant to reading include phonological (Morris et al., 1998; Wagner & Torgesen, 1987; Wagner, Torgesen, & Rashotte, 1979); orthographic (Berninger, 1994a, 1994b, 1995, 2001a); rapid automatic naming or RAN (Wolf, 1986; Wolf, Bally, & Morris, 1986; Wolf & Bowers, 1999; Wolf, Bowers, & Biddle, 2000); morphological (Nagy, Berninger, Abbott, Vaughan, & Vermeulin, in press); and receptive and expressive language (Catts, 1993).

Progress has also been made in understanding the processes underlying handwriting, spelling and composing (for recent review, see Berninger & Amtmann, 2003; Berninger, Abbott, Thomson, & Raskind, 2001; Berninger & Richards, 2002; Wong & Berninger, in press). These processes include orthographic coding in short-term/working memory and long-term memory, motor planning for sequential finger movements, automatic letter production, expressive language, and executive functions, as well as the same phonological and rapid automatic naming processes as for reading.

Screening processing difficulties

Standardized measures now are available to assess these various processes related to assessing and teaching reading, handwriting, spelling, and composing. Rather than waiting for students to fail long enough to meet IQ-achievement discrepancy, schools might develop screening batteries to identify kindergarten, first and second graders who are most in need of early intervention from the following tests developed by researchers: Comprehensive Test of Phonological Processing (CTOPP) (Wagner, Torgesen, & Rashotte, 1999), Test of Word Reading Efficiency (TOWRE) (Torgesen, Wagner, & Rashotte, 1999), Process Assessment of the Learner (PAL) (Berninger, 2001a), Test of Rapid Automatic Naming (Wolf & Denckla, 2004), Clinical Evaluation of Language Fundamentals IV (CELF IV) (Seml, Wiig, & Secord, 2004), and Delis Kaplan Executive Function System (DKEFS) (Delis, Kaplan, & Kramer, 2003).

For older students of normal intelligence with persisting reading and writing problems, documenting the presence of a deficit in orthographic, phonological, RAN, fine motor planning, automatic letter writing, expressive language, or executive function might qualify a student for specialized instruction in reading or writing even if the student does not have an ability-achievement discrepancy.

Linking assessment of processes to instruction

To emphasize the research progress that has been made in linking processes of the learner assessed through testing and taught during instruction, we use as an example the research program at the NICHD-funded University of Washington Multidisciplinary Learning Disabilities Center and the Write Stuff Intervention Project. (See Berninger and Abbott , 2003, for the published research supporting each PAL Lesson Set.)

Sound Games (Berninger, 1998a) train the same phonological processes assessed in PAL Syllables and PAL Phonemes (Berninger, 2001a) and result in reading (PAL Lesson Set 1) and spelling (PAL Lesson Set 2) gains – but gains are not as large as when alphabetic principle is also trained.

Looking Games (Berninger, 1998a) train the same orthographic processes as in PAL receptive coding (Berninger, 2001a) and result in reading gains (PAL Lesson Sets 11 and 12) and spelling gains (PAL Lesson Set 8).

The Talking Letters program (all lessons in Berninger & Abbott, 2003) trains correspondences between orthographic units and phonological units in one of two directions-from the orthographic to the phonological (Pseudoword Reading, Berninger, 2001a) and from the phonological to the orthographic (Spell Sounds subtest on WJ III, Woodcock et al., 2001) because the correspondences work differently in reading than spelling (Berninger, 1998b). This training is directed both to automatic and strategic application of alphabet principle to word decoding or spelling in most lessons.

The Handwriting Lessons (Lesson Set 3; Berninger, 1998b) train fine motor planning for letter formation and orthographic processes underlying automatic production of letter forms and result in improved handwriting automaticity (effortless production of legible letters) and composition. Training morphological processes leads to improved reading and normalized brain activation during morphological processing (Aylward et al., 2003; Berninger, Nagy et al., 2003). The oral summarization activities in Lesson Set 2 and the oral reflective discussions in Lesson Sets 6 and 7 train expressive language skills.

The reading comprehension and written composition activities in all lessons have activities for training executive functions for self-regulated learning. Research in the University of Washington group, as well as other research groups (reviewed in Berninger, Dunn et al., in press) is validating a process approach to both assessment and instruction based on behavioral improvement and brain evidence of normalization.

Response to intervention

At an NICHD-sponsored conference in 1993 on improving measurement of learning disabilities, the construct of response to intervention (RTI) emerged as a viable way of defining opportunity to learn in an era when decoding was not being taught explicitly to many beginning readers (Berninger & Abbott, 1994; Vellutino, Scanlon, & Tanzman, 1994). Subsequently many investigators used growth curve analyses to chart individual students’ progress over time when an intervention was implemented. This research differentiated at-risk children who needed

  • Only short-term intervention
  • More systematic, sustained intervention over time
  • Alternative, specialized instruction taking into account the processes that might be contributing to an individual student’s slow or nonexistent progress (e.g., Berninger et al., 2002; Torgesen, Wagner, Rashotte, Rose et al., 1999; Vellutino, Scanlan, & Lyon, 2000; Vellutino et al., 1996)

Systematic procedures have been developed and proposed for using response to intervention as an approach to early identification and intervention with students and in problem-solving consultation with teachers (Compton, 2000; Fuchs, Fuchs, & Speece, 2002; Stage, 2001; Vellutino, Scanlon, & Tanzman, 1998) as well as to later identification of students with persisting difficulties for special education (Fuchs et al., 2002). These procedures are already place in school districts in Iowa, Minnesota, Washington State, California and elsewhere.

Different kinds of instruments can be used for progress monitoring, including norm-referenced (e.g., Wechsler Individual Achievement Test, 2nd edition [WIAT II], The Psychological Corporation, 2001; or Woodcock-Johnson III [WJ III], Woodcock, McGrew, & Mather, 2001), criterion-referenced (e.g., Academic Competence Evaluation Systems or ACES, DiPerna & Elliott, 2000; Academic Intervention Monitoring System or AIMS, Elliott, DiPerna, & Shapiro, 2001; and The Texas Primary Reading Inventory, Texas Educational Agency, 2000), and curriculum-based approaches.

Curriculum-based approaches

Curriculum-based assessment

We distinguish between curriculum-based assessment (CBA) and curriculum-based measurement (CBM). CBA evaluates whether all the necessary, developmentally appropriate components of curriculum, from a cognitive, linguistic and neuropsychological perspective, are in place (Berninger, 1998a; Naglieri, 2002; Peverly, 1994; Peverly & Kitzen, 1998). Through teacher-administered probes in each lesson, CBA assesses and plots on growth graphs whether students meet the specific instructional goals in a specific lesson (e.g., automaticity of applying letter-sound correspondences to new pseudoword contexts, Berninger, Abbott, Billingsley, & Nagy, 2001; learning taught and transfer words, Lesson Sets 1, 2, 4, and 5, Berninger & Abbott, 2003; increasing fluency after repeated reading, Ihtop & Ihtop, 1989; Berninger, Abbott, Abbott, Graham, & Richards, 2001; Lesson Set 12, Berninger & Abbott, 2003; and applying taught phonological units in a particular pseudoword context to new pseudoword contexts, thereby demonstrating transfer of knowledge from familiar words to unfamiliar words, Berninger, Nagy et al., 2003).

Curriculum-based measurement

In contrast, CBM evaluates the rate of performance of target behaviors and progress on these behaviors over time. Currently many state guidelines for qualifying students for special education rely only on standardized measures of accuracy of performance, and CBM has provided an important, rate-based metric. Recent research has documented that (a) some forms of reading disability may only be detected with rate-based measures, and (b) the automaticity of single word recognition and fluency of oral reading of text are critical for the reading development of all students (see Kame’enui & Simmons, 2001, Lovett, 1987; Wolf, 2001). Hasbrouck has developed innovative ways to integrate CBM rate metrics with criterion-referenced informal reading inventories that pinpoint instructional levels (e.g., Hasbrouch & Tindal, 1992). She uses this integrative approach to evaluate effectiveness of research-supported interventions in the state of Washington.

Differential diagnosis based on multi-domain and multi-modal assessment

Despite early intervention and curriculum modifications, some children have persisting reading and writing difficulties but not always for the same reasons. Not all children with persisting reading and writing difficulties have

  • Similar profiles of cognitive/memory, language, motor, attention/executive function and social/emotional functioning
  • Comparable school and home environments
  • The same prenatal or postnatal developmental histories

Thus, their instruction may need to be individually tailored. Differential diagnosis is important because it leads to different instructional recommendations. However, full-scale IQs are not always as appropriate as factor scores (Prifitera, Weiss, & Saklofske, 1998) in differential diagnosis of learning disabilities, which are defined on the basis of inclusionary criteria based on specific kinds of cognitive abilities.

Examples of differential diagnosis

For example, it is proposed that the verbal comprehension factor on the WISC IV (The Psychological Corporation, 2004) be compared to accuracy and rate measures of single word reading or passage reading and dictated spelling to identify dyslexia-a disorder in which reading and spelling fall below the population mean and below expected level based on verbal comprehension (Berninger, 2001b; Berninger, Abbott, Thompson, & Raskind, 2001).

The perceptual reasoning factor on the WISC IV could be compared to the CELF IV (Delis, Kaplan, & Kramer, 2004) Sentence Formulation subtest, a measure of planning for expressive syntax, and standardized measures of morphological and syntactic awareness to identify language learning disability (Wallach & Butler, 1994)-a disorder in which specific language skills fall below the population mean (but within the lower limits of the normal range) and below expected level based on nonverbal reasoning. Children with language learning disability learn language more slowly during the preschool years and may appear to have normal language skills during the school years unless carefully assessed to document problems in using language to learn academic skills and executive functions related to language use. Unlike dyslexics who tend to have relative strengths in morphology and syntax and relative weaknesses in phonology, the language learning disabled have relative weaknesses in phonology, morphology and syntax.

In contrast, the language skills of those with specific language impairment (SLI) fall outside the normal range (e.g., on the CELF IV expressive language scale) and the substantial language impairment cannot be explained by cognitive ability, as assessed by a measure of nonverbal reasoning.

A multi-axial, multi-domain approach

Different genetic and neurological etiologies may be associated with dyslexia, language learning disability, and specific language impairment (Berninger, in press-a), and these different learning disorders have different treatment needs. Dyslexics benefit from explicit decoding instruction in the context of balanced reading instruction (Berninger, Nagy et al., 2003), but the language learning disabled also need explicit instruction in morphological and syntactic awareness, vocabulary and reading comprehension (Berninger et al., 2002; Wallach & Butler, 1994). Those with SLI need very intensive, systematic and specialized language therapy by a qualified professional.

Quantifying the amount of discrepancy is always arbitrary, and using only discrepancy criteria is always problematic. In the case of dyslexia, underachievement in accuracy or rate of single word reading or spelling or oral reading fluency may be more diagnostic when research-supported language phenotypic markers such as orthographic, phonological and rapid automatic naming deficits co-occur with the underachievement relative to verbal comprehension factor (Berninger et al., 2001).

Berninger (1998a; 2001b) proposed a multi-axial/multi-domain approach to differential diagnosis of specific kinds of learning and developmental disabilities such as dyslexia (Berninger, 2001b), dysgraphia (Berninger, in press-b), language learning disability (Wallach & Butler, 1994), primary language disability or specific language impairment (Catts, 1993), pervasive developmental disorder, and mental retardation. Further longitudinal research is needed on the etiology, treatment, and prognosis for specific academic skills in children with each of these specific learning or developmental disabilities. Research in progress is addressing whether comparison of Listening Comprehension, Oral Expression, Reading Comprehension and Written Expression subtests on the Wechsler Individual Achievement Test-Second Edition (WIAT II) (The Psychological Corporation, 2001) and subtests of the CELF IV (Seml et al., 2004) and DKEFS (Delis, Kaplan, & Kramer, 2003) might contribute to these differential diagnoses.

Mandates versus research

In the meantime, it is very important to make a distinction between federal or state guidelines for qualifying children for special education services and research-supported differential diagnosis (Berninger, 1998a). Many parents are seeking diagnoses, that is, explanations of why their children are having trouble learning, and not just requesting special services. Failure to provide these explanations sometimes leads to adversarial home-school relationships or unreasonable expectations for achievement for some children. For example, the number of children eligible for special education under the category of mental retardation has dropped sharply (Gresham, 2002; Reschly, 2002). This drop may reflect a change in assessment practices and a reluctance to make and share the diagnosis of mental retardation; it does not signal a change in the prevalence of mental retardation in the population. Those children whose intelligence and adaptive functioning fall in the mentally retarded range can usually be taught some reading skills but cannot be expected to read (decode and comprehend) on grade level, as defined by the 50th percentile for age peers or grade-level performance on high-stakes tests.

A call for flexible approaches

Rather than choosing among these four approaches, prereferral and multidisciplinary child study teams might draw flexibly from them, depending on which are most relevant to the child and school setting.

Predicting achievement

One of the mistakes of the past has been to use the IQ score as the single benchmark of expected achievement. Although numerous studies show a reliable, moderate correlation between IQ scores and academic achievement, a growing body of research is identifying language and neuropsychological measures that also should be considered in estimating a reasonable level of academic achievement. The University of Washington research program found that IQ-achievement discrepancy did not predict slope of growth curve in response to early intervention (Stage, Abbott, Jenkins, & Berninger, 2003), but that faster responders (who needed early intervention only in first grade) and slower responders (who needed continued early intervention in second grade) differed significantly in Verbal IQ (verbal comprehension factor) and also in orthographic and rapid automatic naming skills (Berninger et al., 2002).

Enlightened use of intelligence tests

Even though ability-achievement discrepancy alone is not a valid way to qualify children for special education services, it does not mean that (a) intelligence is unrelated to academic achievement or (b) school psychologists should refrain from ever administering intelligence tests. School psychologists need to discover, through research and practice, the most enlightened uses of intelligence tests and to broaden their repertoire to include other tools relevant to linking assessment to instructional intervention. Most importantly, they need to learn more about research-supported instructional interventions to make meaningful recommendations based on the assessments they perform.

Scientifically supported instruction

National Initiatives

Reviews of research. Recently a number of major reviews of effective reading instruction have been conducted. Beginning with the review commissioned by the National Academy of Sciences (chaired by Dr. Catherine Snow) and followed by the National Reading Panel commissioned by the United States Congress (NICHD, 2000), this national effort was supplemented by two meta-analyses-one funded by the NICHD (Steubing, Fletcher, LaDoux, Lyon, et al., 2002) and the other funded by the U.S. Department of Education (Swanson, Hoskyn, & Lee, 1999). Collectively, these reviews on teaching reading point to the importance of explicitly teaching linguistic awareness (phonological, orthographic and morphological awareness of words and their parts), decoding (applying the alphabetic principle of correspondences between 1- and 2- letter units and phonemes), automatic (fast) word recognition, fluency (smooth and coordinated text reading) and reading comprehension (Ehri, Nunes, Stahl, & Willows, 2001; NICHD, 2000; Rayner, Foorman, Perfetti, Pesetsky, & Seidenberg, 2001; Snow, Burns, & Griffin, 1998; Steubing et al., 2002; Swanson et al., 1999).

Effective instruction. A recent review of effective ways to teach reading comprehension found that explicit instruction is needed for developing vocabulary meaning, understanding of sentences, paragraphs and genre-specific discourse structures, and strategies for reading textbooks in the content subjects of the curriculum (Carlisle & Stone, 2002). Explicit instruction directs students’ conscious attention to implicit linguistic knowledge in order to develop both (a) automatic word reading and fluent oral reading of passages, and (b) non-automatic application of strategies and construction of meaning (Berninger, Nagy et al., 2003).

Although national panels have not reviewed writing instruction, recent reviews of writing research (Berninger & Amtmann, 2003; Wong & Berninger, in press) indicate that it should include instruction aimed at (a) legibility and automaticity of letter production; (b) spelling; and (c) composition (word and text fluency; syntactic awareness in sentence construction; genre-specific discourse structures; planning processes; and reviewing and revising processes). Implementing research-supported instructional approaches for literacy in the general education program reduces the number of students needing supplemental early intervention in general education as well as specialized instruction in special education (Foorman, 2003).

Linking assessment to instruction

Neither IQ tests (Reschly & Grimes, 1995) nor traditional academic achievement tests are designed to generate instructional interventions for academic skills. To remedy this situation, a programmatic line of research was begun in 1983 to identify the processes of the learner that influence a child’s ability to learn to read and write and also translate directly into instructional practices for improving reading and writing. Since 1989 this research has been funded by the National Institute of Child Health and Human Development (NICHD).

PAL model of assessment and intervention. The results of the assessment and intervention research are disseminated through the Process Assessment of the Learner (PAL) Test Battery (Berninger, 2001a), Intervention Guides (Berninger, 1998a) and Research-Supported Lesson Plans (Berninger & Abbott, 2003). The second author was the project director for the PAL Intervention Guides. These Guides provide a conceptual model (based on cognitive, linguistic, and neuroscience research reviewed in Berninger & Richards, 2002) for planning instruction and evaluating student progress on the basis of teacher-provided instructional components and student-mediated processing components that are necessary to create functional reading and writing systems. Functional systems are created when knowledge is coordinated rather than fragmented and children begin to think of themselves as readers and writers who choose to read and write in their leisure time and do not avoid or fear school assignments that require reading and writing.

The conceptual model has links to both

  • Planning forms organized by the necessary instructional components and processes of the learner
  • Commercially available instructional resources school psychologists can share with teachers that are organized by the same instructional components/learners’ processes

The list of instructional resources was updated in the PAL Reading and Writing Lessons.

Translating research to practice. The third author was the project director for the PAL Reading and Writing Lessons, which translated research studies into lesson plans that teachers, psychologists, and speech and language pathologists can use with students in grades one to six. These lessons are based on research designs that randomly assigned students or schools to treatment and control groups (Lesson Sets 1 to 10) or to alternative treatments (Lesson Sets 12 and 13) or that use the learning sciences methodology of design experiments (Brown, 1992) to achieve desired outcomes (Lesson Sets 11, 13, and 14). Translating scientific instructional research to educational practice is challenging (Kratochwill & Stoibert, 2000). To deal with some of these challenges, each PAL lesson contains recommendations for (a) progress monitoring to evaluate whether the lesson is achieving the intended instructional goal(s) for individual students and (b) optional adaptations to modify the lesson for those individual students who may not be making progress.

Other publishers (e.g., Paul H. Brookes, Sopris West, Learning by Design and Educators Publishing Service) are also facilitating the translation of research-supported instruction to teacher-friendly instructional materials. Two are noteworthy because they are relevant to teaching students in the upper elementary, middle school and high school years to read and spell longer, morphologically, orthographically and phonologically complex words (Henry, 2003; Masterson, Apel, & Wasowicz, 2002). Another one is valuable because it reviews instructional techniques to help students in these same grade levels expand vocabulary knowledge, understand sentence syntax, paragraph structure and discourse organization, and apply content-specific strategies for reading texts in academic subjects of the curriculum (Carlisle, 2000).

Three-tier model of service delivery for bridging general and special education

The three-tier model for meeting the literacy needs of all students in both general and special education (Berninger, 2002b; Berninger, Stage, Smith, & Hildebrand, 2001) is based on the research of two school psychology faculty at the University of Washington: (a) Berninger’s NICHD-funded research programs on early intervention to prevent reading and writing disabilities, supplementary instruction to help at-risk students meet high-stakes standards, and treatment of severe reading and writing disorders; and (b) Stage’s work with the Washington State Alternative Delivery Project, which found that student progress-monitoring during early intervention and collaborative problem-solving between school psychologists and classroom teachers reduced the number of students requiring special education in the later grades.


The first tier provides early intervention for at-risk readers and at-risk writers. The second tier focuses on student progress-monitoring and curriculum-modification, when necessary, to ensure that students meet high stakes standards in reading and writing. The third tier involves diagnosis to (a) determine why the student has persisting reading and/or writing problems and (b) plan long-term provision of specially designed, individually tailored, explicit, systematic and intensive instruction. The assumption is that with more resources directed toward the first tier, fewer are needed for the third tier.

Both assessment and intervention play a role in each tier. Assessment is needed to identify students needing service at a particular tier and to evaluate their progress in response to the services. However, in the remainder of this article we focus on research-supported instructional approaches at each tier because school psychologists are being asked to become more involved in the design and delivery (or supervision of paraprofessionals in delivery) of instructional interventions relevant to the federal initiatives to “leave no child behind,” to increase the number of students meeting high-stakes standards, and to use research-supported instructional approaches in special education.

Tier 1

A partnership initiated by the Los Angeles Unified School District with the University of Washington is evaluating the use of tier 1 intervention (and associated assessment) (Dunn, 2002). Preliminary evidence shows that combining early intervention in reading (Lesson Set 1, Berninger & Abbott, 2003) and writing (Lesson Set 3, Berninger & Abbott, 2003) is more effective than focusing on reading alone.

Tier 1 intervention appears to work for students whose first language is not English or who do not hear English or mainstream English spoken at home (Quiroga, Lemos-Britton, Mostafapour, & Berninger, 2002; Berninger, Dunn, Lin, & Shimada, in press) as well as for those whose first language is English (Berninger, Abbott et al., 2002). The American School in London (under the leadership of Ted Alper, Professor of School Psychology at For the tier 2 at-risk readers and writers, when orthographic, phonological, morphological and oral vocabulary factors were used to predict a variety of literacy goals, each factor explained some unique variance. However, the orthographic factor explained the most unique variance (significant paths to the most literacy goals) (Nagy et al., in press), which is not totally surprising since what is unique to reading and writing compared to listening and speaking is the orthographic word form. Helping children develop orthographic awareness of letter units that correspond to phonemes or morphemes helps them combine parts of orthographic, phonological and morphological word forms to read and spell words (Berninger & Richards, 2002; Berninger, Abbott et al., 2001; Nagy et al., in press). At Tier 2, combining explicit word recognition and comprehension instruction is more effective than word recognition or comprehension alone (Berninger, Vermeulin et al., 2003), and combining explicit spelling and composition instruction is more efficient than spelling or composition training alone (Berninger, Vaughan et al., 2002). Tier 3 A programmatic line of research combining instruction and brain imaging at the University of Washington is showing that tier 3 intervention (Berninger, Nagy et al., 2003) is effective in treating dyslexia, a genetically based reading disorder (Chapman et al., 2003; Wijsman et al., 2000), and in normalizing brain activation for phonological (Aylward et al., 2003; Richards et al., 2002) and morphological (Aylward et al., 2003) processing. Lesson Set 15 (Berninger & Abbott, 2003) was used in the first study showing brain changes in phonological processing following reading instruction (Richards et al., 2000). Because genetic research is showing that CTOPP nonword repetition (Wagner et al., 1999) is a major source of genetic constraint (Wijsman et al., 2000) and TOWRE rate of phonological decoding (Torgesen et al., 1999) is another major source of genetic constraint (Chapman et al, 2003), tier 3 interventions should focus on developing precise representations of the phonological word form and awareness of its parts (see Lesson Sets 11, 12, and 15 in Berninger & Abbott, 2003) and automatic application of letter-sound correspondences to pseudowords (see Lesson Set 11 Berninger & Abbott, 2003; Berninger, Nagy et al., 2003).

These examples show how genetic research validates the relevant processes that lead directly to the design of instructional interventions and evaluation of their effectiveness based on behavioral and brain changes. These recent research developments that integrate biological and instructional research to study nature-nurture interactions may radically alter the practice of school psychology later in the 21st century.

Future implementation: Systems level change

Implementation of the three-tier model requires systems-level change in delivery of services rather than trying to serve one child at a time (Shapiro, 2000). Denton and Fletcher (2003) also describe the benefits of scaling up for systems-level intervention. Such scale-up may include another three-tier model for alternative delivery of general and special education instructional services (Denton & Mathes, 2003).

In the Denton and Mathes model, the first tier implements research-supported instructional programs in the general education program; the second tier provides supplementary instruction in the general education program; and the third tier offers differentiated, specialized instruction in special education. One way school psychologists in schools with reading achievement in the lowest quartile can implement systems-wide change is to write proposals to seek funding for Early Reading Initiative grants that support research-based efforts to improve reading achievement at the first and second tiers of the Denton and Mathes three-tier model. Helping every child become a reader and writer to the best of his or her ability will be a multidisciplinary team effort that includes children, parents, and school- and university-based professionals working together.


This article has been posted with the permission of NASP as part of the NASP-Reading Rockets Partnership. NASP retains the copyright of these materials. All reprint or use permission should be directed to NASP via

For any reprint requests, please contact the author or publisher listed.