Reprinted with permission from Paul H. Brookes Publishing This chapter originally appeared in Frames of references for the assessment of learning disabilities. New views on measurement Issues, Edited by G. R. Lyon Copyright 1994

Writing and speaking are the basic mechanisms for communication. Although extensive efforts have been devoted to increasing our understanding of spoken output, it has been only in the past decade that an increased interest in the study of written expression has begun to receive the attention of researchers, clinicians, and school personnel (Newcomer & Barenbaum, 1991). Theoretical models for disorders in the writing process have been discussed since the 1960s (e.g., Johnson & Myklebust, 1967; Luria, 1970; Myklebust, 1965), and diagnosis and treatment of writing disorders have been legislated by the U.S. Office of Education since the mid- 1970s (PL 94-142; U.S. Office of Education, 1976), but the study of the writing process has been significantly neglected in regular as well as special education curricula (Bridge & Hiebert, 1985; Roit & McKenzie, 1985). In fact, the study of written expression, particularly with respect to its development and behavioral expression (i.e., impaired as well as gifted performance), has lagged well behind the investigation of other academic domains (e.g., reading and, to some extent, mathematics).

For example, Leinhardt, Zigmond, and Cooley (1980) reported on the relatively small amount of instructional time devoted to educating students about all aspects of the writing process, with the time that was spent on writing largely being devoted to copying tasks. Furthermore, Schenck (1981) noted that it was rare for individualized education programs (IEPs) to include specific objectives to enhance the writing skills of children receiving special education services. Similarly, Christenson, Thurlow, Ysseldyke, and McVicar (1989) reported the following results in support of Leinhardt et al. (1980) and Schenck (1981):

There were no significant differences between the average time spent in general and special education classrooms on formal writing instruction. Students with learning disabilities and behavioral emotional handicaps spent significantly less time writing in content areas compared to time spent in formal writing instruction. Students with learning disabilities and those with mental retardation spent significantly more time receiving written language instruction in the special education setting. The most recent report from the National Assessment of Educational Progress (NAEP, 1990) concerning the nation's writing skills suggested that high numbers of students in the fourth and eighth grades cannot write more than informative essays. A cursory review of the writing literature, however, does suggest that the study of written language expression has surfaced with a higher degree of sophistication than emergent research in other academic areas. This research has been more detailed in its exploration of the subcomponents of the writing process as opposed to conceptualizing writing as a unitary process. It also has been more focused on assessment-treatment linkages as opposed to solely diagnostic endeavors. These factors would indicate that our knowledge of written language expression will increase dramatically over the next decade.

This chapter addresses several key issues in the research and clinical practices pertaining to the assessment of written language disorders. After an initial review of historical foundations, definitional issues, and available epiderniologcal data, a number of assessment issues are discussed. This latter section includes a discussion regarding the conditions presumed to be important for adequate written language production, what models have been proposed to date, and the related issues of neuropsychological underpinnings, developmental precursors and predictors, and possible comorbid conditions. Representative formal and informal diagnostic instruments are described as well. The chapter concludes with a discussion of future needs and directions for clinical and research efforts for the assessment of written language expression.

Historical Foundations
Historically, the term agraphia has been used to describe disorders of written language expression. Well over 100 years ago, Ogle (1867) used this term to describe the relationship between aphasia and agraphia, although he also noted that these two disorders were dissociable. In contrast, several decades later Lichtheim (1885) suggested that disorders of writing were generally the same as disorders of speech. Lichtheim asserted that the acquisition of writing and spelling was superimposed on speech and, consequently, they shared the same neuroanatomical location. Similar conceptualizations emerged during the 1900s (Goldstein, 1948; Head, 1926; Nielson, 1946), largely based on clinical observations and case study methodology, with much of the discussion centering on the issue of association and dissociation of written and oral expression.

To date, much of the research pertaining to agraphia has continued to employ case study methodology (Basso, Taborelli, & Vignolo, 1978) and/or has evolved from examination of individuals with acquired brain injury (Beauvois & D6rouesn6, 1981). Newcomer, Nodine, and Barenbaum (1988) noted in their review of the written language literature that there is a relative dearth of current neuropsychological information pertaining to disorders of written language in children and adolescents, and that definitional issues remain unresolved.

Prior to 1980, agraphia was viewed by most professionals as a "poor relation of aphasia" (Shallice, 1988). Given this association, writing was conceptualized as a secondary skill having little theoretical interest. Consequently, most of the research efforts were directed to understanding the primary language and/or motor-based disorders that were contributing to the writing problem. For example, Chedru and Geschwind (1972) noted that writing problems were largely manifestations of attentional difficulties. This view held that writing was affected because it was not a set of skills that was over learned. Relatedly, these views have contributed to the conceptualization that delays in written expression invariably are associated with reading development and/or are governed by similar metacognitive processes (e.g., planning, self-monitoring, self-evaluation, self-modification) (Wong, 199 1). In general, these perspectives have contributed to the slow acknowledgment of the need to define written language expression in any detail.

Definitional Issues
Since the 1980s, writing has been conceptualized as a problem-solving process whereby authors attempt to produce visible, understandable, and legible language reflecting their declarative knowledge (i.e., knowledge of the topic). The quality of written products is increased or constrained by a multitude of documented factors, such as an author's declarative knowledge, procedural knowledge (i.e., knowledge of the writing process and various strategies to produce different genres of text), and conditional knowledge (i.e., understanding of which writing strategy to use to produce a specific genre for a particular audience, or an audience with unknown needs and expectations) (Gagne, 1985). Less well documented are neuropsychological factors (e.g., attention, language, memory, neuromotor functions, social cognition); personality factors (e.g., motivation, self-esteem); and other endogenous and exogenous factors contributing to the writing process (e.g., teacher-student relationships, amount of writing instruction, teacher knowledge of the writing process). Given the complexity of the writing process, it is not surprising that there have been significant difficulties in establishing a clear operational definition of written language expression. For example, emergent research on written language indicates that most, if not all, students with teaming disabilities have problems with some component of writing, whether it is handwriting, spelling, written syntax, vocabulary, or written discourse. It is important to note, however, that not all aspects of written expression need be impaired in any one child. Although the specific components of written language expression continue to be debated, it seems clear that deriving an operational definition may not be an easy goal to achieve.
Recent definitions pertaining to disorders of written language expression have not proven to be useful to researchers, diagnosticians, or teachers. For example, the teaming disabilities definition offered by the Interagency Committee on Learning Disabilities (ICLD, 1987) notes a multitude of variables associated with teaming problems, including writing, but it does little to operationalize any of the academic difficulties experienced by children. Furthermore, although the most recent revision of the Diagnostic and Statistical Manual (DSM-III-R; American Psychiatric Association, 1987) should be commended for including a category on Developmental Expressive Writing Disorder, this category is poorly defined, conceptualizes writing problems as a unitary construct, and, in general, is not operationalized to a degree to be clinically useful. Although a number of suggestions have been offered with respect to improving the clinical utility of this category (Gregg, 1992), the major criterion for this category continues to depend on the use of ability-achievement discrepancy models. The problems with these models notwithstanding (Kavale, 1987), the availability of reliable and valid assessment tools in the written language domain remains sparse. Finally, one of the criteria in the DSM-111-R specifies that Developmental Expressive Writing Disorder not be due to an underlying neurological disorder or dysfunction, a concept that is in direct contrast to current perspectives on teaming problems in general. Table I lists the DSM-111-R diagnostic criteria for Developmental Expressive Writing Disorder, and it should be noted that these criteria remain largely unchanged for the proposed DSM-IV.

Table 1. Diagnostic criteria for Developmental Expressive Writing Disorder as defined by the Diagnostic and Statistical Manual (3rd ed., rev.)
Writing skills, as measured by a standardized, individually administered test, are markedly below the expected level given the person's schooling and intellectual capacity (as determined by an individually administered IQ test). The disturbance in A significantly interferes with academic achievement or activities of daily living requiring the composition of written texts (spelling words and expressing thoughts in grammatically correct sentences and organized paragraphs). Not due to a defect in visual or hearing acuity or a neurologic disorder. From American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd ed., rev., p.43). Washington, DC: American Psychiatric Association, reprinted by permission.

Ability-Achievement Discrepancy Models
As previously noted, one way to operationalize written language problems has involved ability-achievement discrepancy. The litany of statistical problems associated with ability-achievement models for defining leaming problems has been discussed elsewhere (e.g., Kavale, 1987) and will not be reiterated here in any detail. Suffice it to say, reading its application to written language expression likely will result in confusion similar to that apparent in the reading literature (e.g., Stanovich, 1992). Moreover, as noted with respect to the DSM-111-R definition, the "achievement" component of the discrepancy model has not been operationalized adequately in the writing domain. A single score designed to capture the multiple components of the writing process likely will be of little use in differential diagnosis, and it certainly will not be helpful in designing treatment programs.

The "ability" part of the model has similar problems. For example, the use of "IQ" to determine an expected level of functioning in any academic domain (i.e., set limits) may not be appropriate, and this argument has been cogently put forth by Siegel (1989, 1990) with respect to academic achievement in general. Indeed, the very abilities that are tapped by most "ability" measures are those that typically are deficient in children with learning problems. Consequently, the use of this procedure will lower the "expectancy" level and minimize the chances of obtaining an IQ-achievement discrepancy. In her arguments, Siegel (1989) also highlighted the bidirectional nature of "ability" and achievement factors (i.e., each one can influence the other). Of importance to this chapter, it remains uncertain how IQ will predict written language functioning.

Relatedly, as Stanovich (1986) noted, the discrepancy model does not take into account Matthew effects. In the example of writing, Matthew effects would be illustrated by the findings that good writers likely will become better at written expression and related neuropsychological functions simply by engaging in writing tasks, whereas poor writers will become less proficient over time as a result of disuse and/or slower progress. Furthermore, Lyon (1987) asserted that nonverbal measures of IQ, and even global measures of ability, "bear little psychometric relationship to measures of academic achievement, particularly in the oral language, reading, and written language domains" (p. 78), and the use of these types of measures has contributed to the confusion in the study of learning problems in general. In addition, the discrepancy model appears to have little relationship to the ability to profit from remediation efforts and, intuitively, adds little to specific programming for individual cases.

Alternative Conceptualizations
In response to the many methodological and pragmatic criticisms of the ability achievement discrepancy model, particularly the use of "IQ" as the sine qua non of ability, attempts at modifying this model have been proposed. For example, in the reading domain, Stanovich (1992) suggested that an ability-achievement discrepancy could be operationalized as the difference between listening comprehension (i.e., ability) and reading comprehension (i.e., achievement). This procedure would alleviate "some of the psychometric, conceptual and logical paradoxes connected with the use of intelligence measures" (p. 198). Similarly, within the written language expression domain, one could characterize an ability-achievement discrepancy as the difference between oral expression versus written expression. These types of creative, and potentially more clinically useful, solutions to defining specific learning problems appear to hold promise for diagnosis and, perhaps, treatment; however, guarded optimism is advised because the reduction of any complex learning process, such as writing, to a simplistic formula will need to have sufficient breadth to be effective.

Another modification of the ability-achievement discrepancy model has been to use an "absolute criterion" of low achievement performance. In this approach, an academic deficit is operationalized as one that falls significantly below expected normative standards. For example, any academic performance that falls I SD or more below the mean (i.e., 16th percentile) could be operationalized as a learning problem. This approach has the advantage of capturing all of the individuals with lower functioning, as defined by an arbitrary cutoff; however, it is unlikely to identify students with higher functioning capacity who are not working up to their abilities within a certain academic domain. This latter concern may be particularly crucial with respect to the identification of students with written language problems.

As a compromise to this dilemma, Berninger, Hart, Abbott, and Karovsky (1992) used both the absolute criterion method and an ability-achievement discrepancy method to identify children with reading and writing problems. These investigators argued for distinctions among groups with low functioning (i.e., the lowest 5% in the normal distribution), underachievement (i.e., ability-achievement discrepancy), and learning disabilities (i.e., with both low functioning and underachievement) in the conceptualization and definition of learning problems. This type of approach would allow clinicians and researchers to capture the children experiencing learning problems as well as those children who may not be working at a level commensurate with their abilities. Berninger, Hart, Abbott, and Karovsky (1992) advocated a two-stage model of assessment. The first stage would employ an absolute criterion approach and would assist in the identification of children in need of early intervention. When necessary, a second stage could be invoked whereby an ability-achievement approach could be used to facilitate placement in special education programs.

It is clear from the above discussion that we are far from establishing exact criteria defining disorders of written expression, not to mention learning problems, in general. Several creative efforts have been offered in this regard, but the actual reliability, validity, and clinical utility of these proposed operational criteria and approaches will require direct comparison and application to children with written language problems. The complexities inherent within the domain of written expression will challenge many of these creative options, and future research should address these definitional issues in their own right.

Epidemiological Findings
In contrast to epidemiological findings for dyslexia, there have been few studies directly devoted to characterizing the epidemiology of written language expression. Many of the reasons responsible for the lag in the study of written language also have contributed to the lack of epidemiological information. Basso et al. (1978) did note that acquired, isolated disorders of writing occurred very rarely, with their initial findings suggesting a rate of I in every 250 subjects. One might suspect, however, this estimated prevalence rate to be unusually low, particularly given the higher rates typically cited for reading disorders and learning disabilities in general (i.e, anywhere from 2% to 15%). Furthermore, estimates of the prevalence of developmental language disorders have ranged from 8% to 15% in preschool children (Beitchman, Nair, Ferguson, & Patel, 1986). Likewise, it has been shown that many elementary school-age children have histories of developmental language disorder, although the prevalence figures vary (e.g., Aram & Nation, 1980; Maxwell & Wallach, 1984). Taken together, these prevalence rates suggest that disorders of written language would be at least at similar levels, or perhaps higher, given the number of children who may be "falling through the cracks" of our educational system (e.g., "garden-variety poor readers," "gray-area kids") and/or who exhibit mild difficulties in language-based functions. In fact, Lerner (1976) asserted that "poor facility in expressing thoughts through written language is probably the most prevalent disability of the communication skills" (p. 20).

Obviously, prevalence rates, race and gender ratios, and socioeconomic factors all will be influenced by the definition employed for any given disorder and by where the cutoff lines are drawn. However, few studies of these variables have been conducted within the written language domain. One study did show differential gender effects (Berninger & Fuller, in press), but, as expected, this relationship varied with the specific criterion that was used. When an absolute criterion of identification was employed, more boys were identified as having impairments in the written language domain than girls-a finding consistent with evidence on children with developmental language disorders (Stark, Tallal, & Mellits, 1983); however, there was no difference between the sexes when an ability-achievement discrepancy was used.

To begin to address this question, we gathered preliminary population-based data from an entire middle school in Troy, Alabama (N = 531). Troy has a population of approximately 30,000 people and is located in the southeast part of Alabama, about 50 miles south of the state capital of Montgomery. The middle school population was relatively equal in terms of race and gender distribution, and a wide range of socioeconomic strata were represented. The preliminary data comprised a random selection of approximately 30% of the middle school population (n = 149) and included an equal number of boys and girls as well as an equal number of whites and nonwhites.

These preliminary findings will continue to be reevaluated as the remainder of the population is added into the analyses. Furthermore, specific writing criteria will be established so that the issue of prevalence can be addressed more appropriately. The development of these independent criteria also will permit separation of the data from any psychometric concerns that may be inherent to the TOWL-2 (e.g., problems with representative normative data, reliability, validity). The development of adequate prevalence information will be essential to the study of children with written language disorders.

Using the Spontaneous Writing Scale from the Test of Written Language-2 (TOWL-2; Hammill & Larsen, 1988), preliminary findings revealed an alarming number of students falling at least I SD below the mean on the TOWL-2 subtests (score < 7), as well as on the overall score (score < 85) (Table 2). In general based on a normal distribution, the initial rates were approximately double the expected number, ranging from about 34% on Contextual Spelling and Contextual Style to 47% on Thematic Maturity. Furthermore, there also was a large percentage of students falling at least 2 SD below the mean across all of the measures of the TOWL-2. In contrast, however, ratings of handwriting legibility appeared to fall within normal limits, although nearly one half of the sample printed their compositions when they were asked to use cursive. There was a trend for a greater number of nonwhites to evidence writing problems on the TOWL-2 Spontaneous Writing Quotient, at a ratio of nearly 2: 1. Males also showed a slightly greater rate of impairment, approximating a rate of about 1.5:1.

These preliminary findings will continue to be reevaluated as the remainder of the population is added into the analyses. Furthermore, specific writing criteria will be established so that the issue of prevalence can be addressed more appropriately. The development of these independent criteria also will permit separation of the data from any psychometric concerns that may be inherent to the TOWL-2 (e.g., problems with representative normative data, reliability, validity). The development of adequate prevalence information will be essential to the study of children with written language disorders.

Table 2. Percentage of writing problems on the Test of Written Language-2 Spontaneous Writing Sample using an absolute criterion for a population-based sample of children and adolescents (n = 149)
TOWL-2 domain <2 SD <1 SD Average >1 SD >2 SD
Thematic Maturity 14.8% 47.0% 38.9% 14.1% 3.3%
Contextual Vocabulary 14.8% 39.6% 45.0% 15.4% 2.6%
Syntactic Maturity 17.4% 43.6% 35.7% 20.7% 6.0%
Contextual Spelling 14.8% 34.2% 46.4% 19.4% 8.1%
Contextual Style 10.7% 34.2% 52.3% 13.5% 4.1%
Handwriting 4.0% 16.8% 47.0% 36.2% 4.0%
Writing Quotient 18.1% 43.0% 41.6% 15.4% 4.0%

Assessment Issues and Factors
For any assessment, there are a number of conditions that should help to optimize the accuracy and utility of the findings. The most obvious of these conditions are ones that are intrinsic to the individual, but extrinsic variables also appear to be important for the study of any academic function. Furthermore, the characteristics of the measurement tools also should be evaluated in order to provide for an optimal assessment.

Endogenous Factors
Endogenous factors are variables that largely are intrinsic to an individual. Contemporary theories of human cognition utilize endogenous factors such as attention, memory, language, cognitive strategies, metacognitive awareness, and motivation for the purpose of explaining individual differences among learners on various tasks (Resnick, 1985). For example, students must attend to instruction and academic activities, manage active working memory, comprehend verbal and written instruction, express knowledge and ideas, establish and modify goals, and monitor and modify their behaviors (Bandura, 1986; Brophy, 1983; Rosenthal & Zimmerman, 1978; Zimmerman, 1990). Historically, research has been devoted to investigating the psychoeducational and neuropsychological factors inherent in the writing process (Englert et al., 1988; Wong, Wong, & Blankenship, 1989), particularly with respect to identifying underlying components, developmental precursors, and pathognomonic problems in the writing process (e.g., reversals).

In comparison, Hayes (1990) suggested that factors that seem to be related to writing, such as goal setting and task definition, an individual's motivation for writing, and related neurodevelopmental factors, have not been studied as systernatically. The limited research that has been conducted in each of these areas suggests that these endogenous factors contribute significantly to the quality of written output. Furthermore, research using cognitive theories of motivation (e.g., expectation for success X the value of a successful outcome, attribution theory, self-efficacy for classroom learning) would predict differential performance as a function of one's expectancies for success or failure. Likewise, an individual's self-concept, medical conditions, physical handicaps, and more blatant psychiatric manifestations all can interfere with the reliability and validity of assessment findings.

Exogenous Factors
In contrast, relatively fewer studies have investigated the impact of external or exogenous variables on the development of students' knowledge of the writing process and skills related to written language. For example, school-related factors, such as classroom environment variables (e.g., independent versus cooperative learning situations, direct instruction versus time spent writing in content areas), teacher characteristics and instructional style (e.g., attributions for students' success and failure on academic tasks requiring written products; self-efficacy for teaching writing in general and to problem learners in particular), specific teacher-student relationships, class size, teacher preparation and subsequent knowledge about the writing process, and the amount of time spent teaching writing all would appear to be crucial to the ultimate manifestation of a child's written output (Lyon & Moats, 1989). Other exogenous variables such as socioeconomic status, family constellation and stability, and family attitudes toward education and literacy likely will be significant contributors to the quality of the writing process.

In comparison to studying the natural evolution of writing skills, researchers have devoted much effort to instructional variables that promote expertise in writing. In general, intervention research attempts to promote writing skills by altering the demands of the writing process through use of computers (MacArthur & Graham, 1987; MacArthur & Schwartz, 1990, in press), by instruction via expert scaffolding (Englert, 1990; MacArthur, Schwartz, & Graham, 1991; Montague, Graves, & Leavell, 1991), or by direct cognitive and metacognitive strategy instruction (Graham & Harris, 1989). Given these emergent research findings, it is clear that assessment of exogenous variables will be crucial to the diagnosis of writing problems as well as contributing to specific assessment-treatment linkage paradigms.

Instrument Factors
A third general condition for reliable and valid assessment involves the variance that can be associated with any specific diagnostic instrument. In addition to the importance of ensuring satisfactory psychometric standards (i.e., reliability, validity), the choice of instruments also becomes important for operationalizing the assessment process. For example, the choice of instruments may be conducted to address diagnostic issues that are relative to normative as well as ipsative standards, and these instruments may yield information related to components of the writing process and/or a total rating or score. Subsequently, another instrument may be selected that is more dynamic and qualitative in its structure so as to facilitate treatment monitoring and progress. Given the complexities of the written language process, it would seem that there should be an integration of these two kinds of assessment tools, but the variance associated with the instrument(s) selected ultimately will influence the results.

Summary
Any general assessment model should attempt to take into account a variety of factors and, consequently, attempt to address the general conditions for assessment. As illustrated in Figure 1, endogenous variables (e.g., neuropsychological functions, social-emotional characteristics), exogenous variables (e.g., family functioning, socioeconomic factors, classroom variables, teacher knowledge of written expression), and instrument variables (e.g., reliability, validity) all are important in beginning to study the writing process. This is especially important given the relative dearth of information concerning the development of normal as well as deficient writing skills. Relatedly, increased knowledge with respect to the unique contributions of these variables, particularly as they may vary in terms of their developmental importance and/or their contributions to specific aspects of the written language process, also will assist professionals in designing comprehensive treatment plans for children and adolescents with written language disorders and dysfunctions, as well as providing challenging programming for children advanced in the written language domain.

Figure 1. General conditions/variables for the assessment of written language expression.
WHAT ARE WE ASSESSING?
Characteristics of "Expert" and "Poor" Writers
Prior to engaging in a discussion of the specific production components involved in written expression, it is important to understand the general characteristics of "expert writers" and "poor writers" in order to set the stage for more detailed assessment of the writing process. These characteristics are important not only for diagnostic procedures but also for beginning to structure treatment programs. A number of investigators have begun to describe the characteristics of "expert" writers (Bereiter, 1980; Burtis, Bereiter, Scardarnalia, & Tetroe, 1983; Fitzgerald, 1987; Flower & Hayes, 198 1; Halliday & Hasan, 1976; Hayes & Flower, 1986; McCutchen & Perfetti, 1982; Scardamalia & Bereiter, 1986; Sommers, 1980). In general, expert writers are goal directed, moving in a recursive fashion from translating ideas to using generated text to plan the next sentence or paragraph. Each of these steps is evaluated with respect to its relevance to meeting proximal or distal goals.

Embedded within this recursive process is a keen orchestration of the expert writers' understanding of the goals of the writing assignment, their knowledge of the topic, and their knowledge of the audience. In addition to this recursive process, expert writers typically generate more ideas and use an abundance of transitional ties (e.g., referential cohesive ties, connective cohesive ties, lexical cohesive ties) to produce more cohesive text and a smooth flow of ideas. During the various phases of the writing task, expert writers are more likely to attend to the communication of meaning rather than to details of form (e.g., spelling, grammar) and are more likely to make substantive revisions to the meaning of their text in an effort to increase clarity.

In contrast, "poor" writers show not only deficits in their strategies during production of written text but also clear problems in simply generating text (Englert, 1990; Englert, Raphael, Anderson, Gregg, & Anthony, 1989; Englert & Thomas, 1987; Graham & Harris, 1987, 1989; Graham, Harris, MacArthur, & Schwartz, 1991; Graham, Harris, & MacArthur, 1990; Wong et al., 1989).

When compared to students achieving normally, poor writers are more likely to have shorter and less "interesting" essays, produce poorly organized text at both the sentence and paragraph levels, and be delayed in their acquisition and/or use of declarative, procedural, and conditional knowledge about writing. Further more, poor writers are less likely to revise spelling, punctuation, grammar, or the substantive nature of their text to increase communication clarity. In general, it seems that poor writers are simply less knowledgeable about the entire writing process.

What we have gained from attempting to understand the characteristics of expert and poor writers has been helpful in establishing assessment-related factors, increasing linkages to treatment, and suggesting specific interventions for the treatment process. It also has helped us to generate theoretical underpinnings for written language expression and to increase our understanding of the skills necessary to engage in different types of writing.

Different Types of Written Language Expression
It is necessary to recognize that there are many different types of writing that can occur, which, in turn, places importance on the kinds of methods and/or prompts that are used to obtain written language samples. There have been a variety of attempts to categorize the different types of written discourse.

Moffett (1986) suggested that writing products could be divided into five types. One type involved imagination and included writing characterized by fiction, plays, and poetry. A second type involved recollection and included tasks such as autobiographies and memoirs. A third type involved notation where listings typically occurred, such as in a diary or a journal. The fourth kind of writing discourse involved cogitation. Examples here included editorials, reviews, and personal essays. The fifth kind was more investigative and included biographies, nonfiction, and research reports. These different kinds of written discourse served as the basis for the Educational Testing Service's (ETS's) introduction of the first statewide assessment of writing abilities in eighth-grade students in California (ETS, 1986).

Another categorization model of different types of writing proposed by Howie (1984) provided four major modes of composition. These included: 1) narration/description, wherein an individual tells a story using characters, plots, and descriptive information; 2) procedural compositions, in which an individual attempts to relate a process step by step using words of transition; 3) time-order exposition, in which the writer presents events in a time sequence, typically designed to solve a problem and/or address cause and effect; and 4) topic exposition, wherein the writer is required to state an argument or topic that requires supporting details and logical connections.

Still another categorization involving five types of written discourse was proposed by D'Angelo (1984). These included expressive, narrative, descriptive, expository, and argumentative prose. In addition, he noted that poetry and drama also should be included in this categorization. Needless to say, all types of discourse tend to be present in most models, with one model being merely a variation of another. Accordingly, discourse models have been criticized as inadequate conceptual frameworks for written expressive language (Lloyd-Jones, 1977), although they still remain useful for setting the parameters of a writing task.

Emergent developmental data at present would suggest that story production tends to lag behind story comprehension, but that by the end of the primary grades virtually all children are capable of producing well-formed narrative compositions (Botvin & Sutton-Smith, 1977; Bracewell, Frederiksen, & Frederiksen, 1982; King & Rentel, 1981; Stein & Trabasso, 1982). Other forms of written discourse, however, have received less examination. Nonetheless, Scardamalia and Bereiter (1986) reported that, in their experience, children age 10 years and older very rarely experience difficulties in their efforts to conform to some type of written discourse. Other investigators, however, have reported significant difficulties experienced by most students on the more substantive requirements for the different types of written language (NAEP, 1980a, 1980b, 1980c). For example, on persuasive writing tasks, it has been noted that 3 1 % of 13-year-old students and 22% of 17-year-old students were rated as being deficient in their attempts to provide support for a position. Similarly, about one third of 17-year-old students were unable to write within a fictional context.

Information pertaining to the type of written discourse, particularly the emergence of some developmental aspects with respect to specific kinds of stimuli, raises the question of what might be the best kind of stimuli to employ when attempting to obtain a representative writing sample. In addition to the decision about what kind of writing to request, should the student be asked to write from prompts tapping one or more of the sensory modalities (e.g., a picture, a verbal description, a smell)? Should the writing assignment be structured or unstructured? How should these terms be operationalized? Or, should one evaluate students' notebooks, homework assignments, or other written assignments? Information related to making these decisions is only beginning to emerge.

For example, when using a picture stimulus to elicit narrative structure, it is suggested that a real photo be used. In addition, the picture should contain certain features to elicit a well-defined narrative structure: At least two characters, with perhaps one potential protagonist and one potential antagonist, although this is not absolutely necessary, depending on the situation depicted. A depiction of some kind of "interesting" or "novel" scene or event (e.g., children lost in a cave). Some kind of potential conflict between the antagonist and protagonist that necessitates that the protagonist engage in a goal-based sequence of events to resolve the conflict, or a scene in which a main character must resolve a problem using some goal-directed sequence of events. For this kind of stimulus, the major purpose is to elicit an expression of some kind of event sequence.

Although one would expect developmental differences in the written expression of narrative structure, the major components of initiating event (e.g., problem or conflict), attempt (e.g., problem resolution attempts), and direct consequence (e.g., result) should be elicited from most competent writers, whereas setting (e.g., location), internal response (e.g., what the character feels/thinks about the situation), and reaction (e.g., how the character feels about the outcome) may be optional. Other investigators have examined selected aspects of the wording and structure of writing prompts such as the influence of syntactic and rhetorical structures on the quality of student writing. For example, Smith et al. (1985) found that topic structure (e.g., open structure versus structured response to one task versus structured response to three tasks) could be designed for specific levels of writing competence. Other investigators have not found significant results in writing quality as related to whether prompts were personalized or neutral (Brossell & Ash, 1984), had high or low structural demands (Greenberg, 1981), or were long or short (Hoetker & Brossell, 1989). Relatedly, Greenberg (198 1) found that error-free writing exerted a more significant influence on raters than did syntactic maturity. Hoetker and Brossell (1989) did note, however, that a shorter or framed writing prompt appeared more desirable because it seemed to allow students, particularly those with writing impairment, more control over their written Output. Although far from conclusive, these initial findings do suggest that the kinds of writing prompts might be important to or influential on the generation of an adequate writing sample; however, this question will require further study.

Phases in Written Output and Proposed Production Components
In addition to the different types of written language expression, investigators have hypothesized that the writing process is governed by three interdependent phases (Flower & Hayes, 1986): planning, translating, and reviewing. Each of these phases also contains elementary mental processes that are thought to be hierarchically organized.

The planning phase is the process whereby writers form an internalized representation of the knowledge they will use to generate text, although it is important to note that these representations do not necessarily have to be in the form of visible language. Planning is hypothesized to be governed by the follow three elementary mental processes: 1) generating ideas (i.e., retrieval of information stored in long-term memory), 2) organizing ideas (i.e., using a meaningful structure to organize ideas), and 3) goal setting (i.e., procedural, substantive, proximal, distal).

The translating phase is the process that the writer employs to put ideas into visible language (i.e., writing). This phase is hypothesized to be governed by the following five language-driven mental processes: syntactic, lexical, motoric, grammatical, and spelling. The final phase involves reviewing, Reviewing is the process whereby writers read what they have written in order to make modifications to existing text and/or to generate new ideas. This phase is hypothesized to be governed by two basic mental processes: 1) evaluating, which involves making decisions about the quality and quantity of text; and 2) revising, which entails making superficial (e.g., spelling errors) as well as substantive modifications to the text.

Superimposed on these compositional phases are a number of production components that comprise written expression. Production components are the basic elements of a specific function. At present, the literature would support the existence of at least three production components (i.e., organization, ideation, and mechanics), although it is likely that there are other specific production components contributing to written language. For example, Gregg (1992) has discussed the specific components of handwriting, spelling, syntax, text organization, and sense of audience. Within the spelling component, Gregg stated that it was important to analyze error patterns, with emphasis being placed on phonological, visual, morphological, spelling, segmentation, and context types of errors. She also identified the production component of written syntax as important to the written language process and described a number of syntactic errors that can be manifested in written text. These included word omissions, word order errors, incorrect verb and pronoun use, word ending errors, lack of punctuation and capitalization, discrepancy between oral and written language, metalinguistic problems, and problems with cohesion.

Gregg further suggested that text organization was a necessary production component of the written language process. Text organization included cohesion and coherence. Cohesion comprises the use of a variety of formal grammatical and lexical devices that operate at both intra- and intersentence levels to link ideas. Coherence, in contrast, involves sentence organization and intra- and interparagraph organization and generally is concerned with the overall organization and flow of the text. Finally, Gregg asserted that sense of audience was a crucial component in writing. Sense of audience implies that the writer has made the necessary adjustments during the writing process to accommodate the intended readership. Flower (1979) and Shaughnessy (1977) both noted that this production component was crucial to the quality of idea elaboration and the sheer number of details presented in written discourse. Other investigators have identified this component as an important factor in how writers organize their written text (Berkenkotter, 198 1; Flower, 1979) and select specific syntactic structures (Crowhurst & Pich6, 1979; Rubin, 1982; Smith & Swan, 1978). Gregg and McAlexander (1989) also noted the importance of a variety of neuropsychological variables (e.g., visual processing, oral expression) in the manifestation of the quality of sense of audience. Consistent with Luria (1980), Gregg (1992) stated that any assessment of written language should include measures of all of the production components across copying, dictation, and spontaneous writing tasks.

Two other models describing the production components of written Ianguage have been derived empirically. Cohn and Isaacson (1984) performed factor analysis on 13 informal measures obtained from student essays and obtained six reliable factors. These factors were fluency, which involved the generation of simple sentences and compositions of gradually increasing length; syntactic maturity, which was manifested in the production of written sentences of increasing complexity; content, which pertained to the organization and originality of the writing product; conventions, or the mechanical components of the writing task (e.g., punctuation); vocabulary; and spelling.

More recently, Tindal and Parker (1991) performed a factor analysis on several quantifiable indices obtained from the compositions of elementary students. These indices included the number of words written, words spelled correctly, words sequenced correctly, words sequenced incorrectly, total number of sequenced words, percentage of words correctly sequenced, story idea, story organization, and conventions. Results of the factor analysis revealed the presence of three factors accounting for 8 1 % of the variance. The first factor, Production and Sequence of Words, accounted for 37% of the variance and was comprised of the number of words written, words spelled correctly, words sequenced correctly, and total number of sequenced words. The second factor, Conventions, included the indices of percentage of words correctly sequenced and a rating of conventions/ mechanics. The second factor accounted for about 26% of the variance. The last factor, Idea/ Cohesion-Organization, accounted for 18% of the variance and included the measures of story idea and organization. This is one of the better conducted studies examining the production components of written expression, although the utility of these findings will require further examination.

Although variants of the above production components have been proposed (Cartwright, 1969; Cooper, 1977; Myklebust, 1965; Parker, Tindal, & Hasbrouck, 1991; Phelps-Gunn & Phelps-Terasaki, 1982; Poteet, 1980; Straw, 1981; E. Weiner, 1980), those asserted by Gregg 1992) and Tindal and Parker 1991, and the phases proposed by Hayes and Flower (1986), generally are well accepted by professionals in the field. These indices also seem to have the advantage of cutting across product as well as process assessment issues. These production components can be seen in Table 3. However, although these production components appear to be conceptually sound, there are few data available pertaining to the underlying structure of written expression. Furthermore, there are few data examining the specific relationships between these production components, particularly because their relationships may change with development. These questions will require further examination, perhaps using more empirically based techniques (e.g., factor analysis).

Table 3. Examples of proposed production components for written language expression
Author(s) Production Components
Flower and Hayes (1981) Planning - Generating ideas, Organizing ideas, Goal setting
Translating - Syntax, Lexical, Motoric, Grammatical, Spelling
Reviewing - Evaluating, Revising
Cohn and Isaacson (1984) Fluency
Syntactic Maturity
Vocabulary
Content and Organization
Conventions
Spelling
Tindal and Parker (1991) Production and Sequence of Words
Conventions
Idea/Cohesion-Organization
Gregg (1992) Handwriting
Spelling
Syntax
Text Organization
Sense of Audience

Proposed Theoretical Models for Written Language Expression
Approximately 30 years ago, Vygotsky (1962) postulated that the complex process of learning to write involved not only the integrity of cognitive abilities but the integration of these abilities into the social context. Similarly, Luria (1980), described the multitude of neuropsychological factors involved in written language expression (e.g., memory, motor). More specifically, Luria (1970) argues that the writing process involves several steps. First, speech is broken down into individual sounds, with each sound being represented by letters, and then the individual letters are integrated to produce the written word. Despite these earlier efforts, there have been few models developed to examine this process, particularly over the span of development. Furthermore, when one extracts the models that deal exclusively with written spelling (e.g., Frith, 1983; Shallice, 1988), few models of written language remain. Examples of these include models proposed by Abbott and Berninger (1992), Ellis (1982), Hayes and Flower (1980), and Roeltgen (1985).

Hayes and Flower Model This model (Hayes& Flower, 1980) perhaps is one of the most widely cited models to date. It maintains the unique feature of establishing much of the vocabulary that tends to be used when discussing the writing process. Figure 2 shows the general structure of the composing process. The main features of this model involve planning, translating, and reviewing functions that are monitored constantly throughout the writing process. There also are reciprocal relationships between variables in the task environment (e.g. the writing task and the generated text at any point in time) and the writer's long-term memory. Being in the tradition of problem-solving models, this model incorporates process as well as product issues, and it has been relatively robust with respect to examining a wide variety of diagnostic and treatment issues in the writing process (Scardamalia & Bereiter, 1986). A variant of this model, which depends largely on the generating component (i.e., knowledge-telling strategy), also has been proposed to account for the difficulties encountered by immature writers.

Figure 2. Hayes and Flower model depicting the phases of the writing process. (From Hayes, J.R., & Flower, L.S. [1980]. Identifying the organization of writing processes. In L.W. Gregg & E.R. Steinberg {Eds.], Cognitive processes in writing [pp. 3-30]. Hillsdale, NJ: Lawrence Erlbaum Associates; reprinted by permission.)

Ellis Model In Ellis' model (1982) the writing process is described as a series of transformations between different types of representations, with the representations being stored temporarily in a variety of memory buffers. As illustrated in Figure 3, this framework is dependent on a recall hierarchy that distinguishes between graphernes (i.e., abstract letters), allographs (i.e., concrete letter formations), and graphic-motor patterns (i.e., the sequence and direction of the written strokes). Once information reaches the latter stage of the hierarchy, neuromotor execution can occur. Written language difficulties would correspond to the different memory buffers, with a significant contribution of this model coming from the work studying spontaneous lapses of recall or "slips of the pen" in normal writers (Hotopf, 1980; Wing & Baddeley, 1980). For example, letter transpositions are assumed to occur while retrieving information from the graphemic buffer; haplographies (i.e., omission of one of a pair of repeated letters and all the letters lying between them) when retrieving information from the allographic buffer; and letter substitutions when retrieving information from the graphic-motor pattern buffer. Although this model makes a strong case for the importance of memory in the writing process, the model remains limited with respect to its exclusion of other kinds of variables (e.g., neuropsychological).

Figure 3. Ellis model depicting the writing process as a series of transformations between different types of representations, with the representations being temporarily stored in a variety of memory buffers. (From Ellis, A. W. [1982]. Spelling and writing [and reading and speaking]. In W.W. Ellis [Ed.], Normality and pathology in cognitive functions [pp. 113-146]. London: Academic Press; reprinted by permission.)

Roeltgen Model This is one of the more complex models depicting the writing process. As can be seen in Figure 4, Roeltgen (1985) not only describes the various neuropsychological variables in this model but also suggests what neuroanatomical location(s) might be affected in written language deficits or dysfunctions. Indeed, some of these hypothesized brain regions have begun to gain some support via neuroimaging studies (Roeltgen & Heilman, 1984). This model of writing illustrates linguistic and motor relationships, as well as hypothesized interactions between visual-spatial processes and motor components. In addition, this model goes a step further by providing a variety of functional systems that can be disrupted, which, in turn, contributes to specific subtypes of written language manifestations (e.g., phonological agraphia). The model is robust enough to begin to embrace the relationships between written language problems and other neuropsychologically based deficits.

Figure 4. Roeltgen model illustrating the various neuropsychological variables and suspected neuroanatomical sites that might be involved in written language dysfunction. (From Roeltgen, D. [1985. Agraphia. In K.M. Heilman & E. Valenstein [Eds.], Clinical neuropsychology [pp. 75-961. New York: Oxford University Press; reprinted by permission.)

Abbott and Berninger Model More recently, Abbott and Berninger (1992) studied the developmental changes assumed to be inherent in spelling, written composition, and handwriting skills. Using a battery of neuropsychological procedures, which included measures of orthographic coding, neuromotor functioning, visual-motor integration, phonological coding, word finding, sentence syntax, verbal intelligence, real word reading, nonword reading, and passage comprehension, these investigators found significant relationships with their spelling, handwriting, and composition outcome measures, although these relationships seemed to be age dependent. For example, as depicted in Figure 5, for written composition outcomes across narrative and expository tasks, only the relationship between reading skills and the composition factor was significant at Grade 1. However, in Grades 2 and 3, this relationship changed to include both reading and oral language/verbal IQ, with the contribution from the latter factor being larger.

This latter model represents one of the more technically sophisticated attempts to address neurcipsychological contributors to an array of written language skills. In addition to showing differential relationships to these selective writing skills, this model also is important with respect to detailing specific developmental precursors of the writing process. In this regard, such a model begins to set the stage for data-based assessment-treatment propositions and strategies.

Developmental Precursors

Johnson and Myklebust (1967) and Myklebust (1965, 1973) discussed a hierarchy of language abilities and stated that the ability to write developmentally follows the abilities to listen, speak, and read. Litowitz (1981) pointed out that subsumed under the need for those basic language capacities are neuropsychological abilities (e.g., selection attention, perception, categorization, memory, problem solving, visual-spatial abilities), which also will have their own developmental trajectories. Intuitively, this suggests that a wide range of neuropsychological profiles (i.e., strengths and weaknesses) may underlie the same achievement outcome (Berninger & Hart, in press). Furthermore, it is likely that these neuropsychological components do not ascend developmentally at similar rates (i.e., developmental dissociations), nor is their development dependent on one another. Given these assumptions and their associated complexity, it is no accident or oversight that specific findings with respect to the developmental precursors of written expression have been limited. Understanding these precursors takes on increased significance when it is realized that many of these problems continue into adulthood (Temple, 1988).

Some initial efforts have been conducted to examine the developmental precursors of written language expression. For example, Berninger and Hart (in press) noted that, when absolute definitional criteria are employed, few of the children in their early elementary school sample showed deficits in handwriting, spelling, or composition; however, disorders of composition did co-occur with handwriting problems. When using relative criteria, however, these investigators found that a specific writing component could be underdeveloped or overdeveloped independently of the other components that were defined in the writing process. A similar finding was observed when a combination of criteria (i.e., absolute and relative) was employed, although the independence of the component writing skills was even more pronounced.

Berninger, Yates, et al. (1992) also conducted a study to examine the nature of specific predictor measures of the writing process in a group of early primary-grade children. Using 300 first-, second-, and third-grade students, Berninger et al. administered a battery of tasks that encompassed neuromotor, orthographic, visual-motor integration, syllable and phoneme segmentation, word finding, sentence syntax, reading, and verbal intelligence functions. Their criterion measures for written expressive language included narrative and expository written products, spelling, and handwriting. Results revealed that automaticity in the production of alphabet letters, rapid coding of orthographic information, and speed of sequential finger movements were the best predictors of handwriting and composition skills. Spelling was predicted by orthographic-phonological mappings and visual-motor integration. In general, these lower level neurodevelopmental variables were related directly to initial writing skills, and they appeared to have their largest impact on the transcription subcomponent of the writing process. Assessment of these lower level tasks appears to be a necessary component for tracking the development of the writing process (Berninger & Rutberg, 1992).

Beminger, Mizokawa, and Bragg (1991) viewed these findings as supportive of their theoretical model pertaining to the developmental constraints on writing acquisition. In general, this model postulates that neuropsychological, linguistic, and higher cognitive constraints may be recursive throughout the development of the writing process, but that each of these constraints may exert relatively more influence at different points in the developmental processes. For example, lower order neurodevelopmental abilities (e.g., automaticity of alpha bet production, speed of sequential finger movements) may influence the ability to transcribe ideas into written products (e.g., letters, words, simple sentences). Once this is accomplished, linguistic factors at the word, sentence, and paragraph levels will facilitate or hinder the written expression process. Finally, as written expression tasks increase in length and complexity, higher order neurocognitive abilities (e.g., planning, translating, revising) come into prominence in producing written output.

Although relatively sparse, emergent information pertinent to the developmental precursors of written language expression appears fruitful with respect to evolving models of this highly complex process. Furthermore, these findings begin to set the stage for assessment and treatment paradigms that would be useful for clinicians and researchers. Although limited to the level of endogenous variables, these beginning efforts are highly significant and noteworthy. Future investigations should attempt to employ not only an array of endogenous variables, but exogenous factors as well, in formulating developmental predictors of written output functions (i.e., deficient as well as advanced performances).

Multidimensional Nature of Writing Process

The writing process always has been conceptualized as a multidimensional phenomenon, going back to the early writings of Ogle (1867). In fact, Goldstein (1948) and Nielson (1946) even proposed tentative classifications for the different kinds of written output difficulties manifested. For example, Nielson's classification included three types of agraphia (i.e., apractic, aphasic, and isolated), whereas Goldstein proposed five types (i.e., poor impulse for writing, impairment of the abstract attitude, ideatoric agraphia, motor agraphia, and agraphia associated with apraxia of the minor hand). Given the models presented to date, the hypothesized production components and phases, and the speculations regarding neurodevelopmental precursors, the stage has been set for a multidimensional conceptualization of written expression.

Moreover, the evidence presented to date would implicate a variety of neuropsychological underpinnings influencing written language output. For example, Gregg (1991, 1992) noted that there are a variety of language-based processes (e.g., phonological disorders, dysnomia), visual-spatial processes, and executive functions (e.g., organization, planning, evaluating) that can impact on written language functioning. The various models proposed for written language (e.g., Ellis, 1982) also suggest the importance of memory, attention, and other neuropsychological functions for competent written language expression. Similarly, the work of Berninger and colleagues has begun to map some of these neuropsychological functions with respect to their unique contributions to the writing process as well as their significance to writing outcomes at specific periods of development.

Although investigators should be cognizant of the principle of noncontingent variation (i.e., specific neuropsychological combinations do not seem to co-occur across children in a consistent manner for the same disorder, as is characteristic of a syndrome; Berninger & Hart, in press), it is clear that multidimensional models for written expressive language may have some merit. As with learning disabilities in general (Hooper & Willis, 1989; Rourke, 1991), subtyping models that explore the multidimensional nature of the writing process may assist in increasing our understanding of the components of this process, the underlying neuropsychological contributors, and how these intertwine to influence treatment planning and efficacy.

Earlier attempts to classify agraphia typically have been based on clinical case analyses of individuals showing more frank neuropathology (Benson, 1979; H6caen & Albert, 1978; Kaplan & Goodglass, 1981), and consequently, have speculated on the neuroanatomical locations involved in each specific writing deficit. Taken together, these efforts have resulted in five different types being proposed. The first subtype is called pure agraphia and entails writing problems without any other significant language problems. The second, aphasic agraphia, involves language-based problems that interfere with written output. The language-based problems can entail difficulty in graphemic production or agrammatism. The third syndrome, agraphia with alexia or parietal agraphia, involves significant deficits in reading and writing in the absence of significant aphasia. The fourth type, apraxic agraphia, is characterized by deficits in the actual formation of the graphernes when writing spontaneously and to dictation. Finally, spatial agraphia involves problems writing on a horizontal line and poor use of space (e.g., letter/word juxtapositions, spacing errors). Although a useful teaching heuristic, particularly for beginning model development, this classification model is hampered by its dependence on clinical case descriptions, not to mention the inordinate amount of overlap that can be seen across these symptoms (Roeltgen, 1985). Furthermore, because these subtypes were based on clinical descriptions of largely adult patients with frank neuropathology, their relevance to children and adolescents with neurodevelopmental phenomena likely is lessened.

Roeltgen (1985) proposed that, instead of classifying patients based on their clinical manifestations, classifications should be driven by neuropsychological functions that are assumed to be disrupted in the writing process. Based on the model discussed earlier, Roeltgen proposed a number of subtypes organized around the linguistic, visual-spatial, and motor components of writing. Specific subtypes include lexical agraphia, phonological agraphia, deep agraphia, semantic agraphia, phonological agraphia with Wernicke aphasia, lexical agraphia with semantic paragraphia, apraxic agraphia, spatial agraphia, and unilateral agraphia. Although this model has the advantage of describing the underlying neuropsychological systems involved in writing difficulties, it also is based on a pathological model of functioning and is derived largely from case study descriptions. Again, these speculations limit the generalizability of such a model to the understanding of normal, advanced, or deficient written language expression in children and adolescents.

More recently, using children with neurodevelopmental disorders, Berninger et al. (1991) hypothesized that one kind of writing disability may involve the inability to translate knowledge of written language into an actual written product, despite the presence of intact functioning in other domains. A second kind of written language deficit would be manifested by slow composing time as opposed to any major inaccuracies in the actual output. This kind of deficit would be contrasted with another in which composing times may be inordinately slow and output inaccurate. Two other kinds of writing difficulties would occur at the word level of analysis, with one kind evidencing a specific deficit in word production only and the other showing deficits in both word decision and word production. Specific treatment suggestions were made by Berninger et al. for each of these kinds of writing problems.

One of the first empirical studies to address the multidimensional nature of written expression in children and adolescents was conducted by Sandier et al. (1992). Based on teacher ratings of writing legibility, mechanics, writing rate, spelling, and written language sophistication, clinic-referred children, ages 9-15, were grouped into those having a writing disorder (n = 99) and those without (n = 63). All subjects received a neurodevelopmental battery designed to screen eight broad areas of functioning: minor neurologic indicators, fine-motor function, language, gross-motor function, temporal-sequential organization, visual processing, selective attention, and memory. Cluster-analytic techniques produced six reliable clusters, four of which comprised the writing disorder group.

The largest group (n = 50) evidenced fine-motor and linguistic deficits, with writing problems characterized by poor phonetic spelling, slow motor output, and problems with mechanics. The second group (n = 35) predominantly manifested visual-spatial deficits, with their writing being characterized by poor legibility and poor spatial organization. In contrast to Group 1, however, these children evidenced adequate spelling and idea generation. Group 3 (n = 9) showed prominent problems with attention and memory. Their writing was characterized by poor spelling with frequent omissions, insertions, and inconsistencies. Legibility, mechanics, and rate generally were intact. The final writing disorder group (n = 11) showed primary deficits in their sequencing, and their written output was characterized by poor automatization of letter production poor mechanics, and decreased legibility. Although this study could be criticize from a number of perspectives (e.g., sampling, assessment procedures), it is important in that it provides further support for the heterogeneous nature of written language dysfunction. These findings complement previous efforts in that they go beyond clinical observations of adults with frank brain involvement and reflect proposed frameworks for neurodevelopmental dysfunctions in written expression in children and adolescents (e.g., Gregg, 1992).

Although several other tentative subtype models of written language expression have been discussed (e.g., Gaddes, 1985; O'Hare & Brown, 1989), particularly with respect to spelling (Finucci, Isaacs, Whitehouse, & Childs, 1982; Frith, 1983; Nelson & Warrington, 1976; Sweeney & Rourke, 1978), it is clear that there may be possible benefits from studying syndromes of written language expression. Although the unit of analysis (i.e., group data versus individual data) debates will continue, increased understanding of the different manifestations of written language functioning and related neuropsychological underpinnings holds the potential to facilitate the development of specific treatment paradigms for children experiencing problems in this domain. It also may suggest methods whereby normal and advanced writers could become more efficient and effective in their written work.

Issues of Comorbidity

Given the hypothesized models and multidimensional aspects of writing, it should not be surprising that this heterogeneous group of disorders would be plagued by a number of co-occurring features. In particular, if writing does indeed represent the last language function to develop (Myklebust, 1973), language problems and associated academic problems may be pervasive throughout a child's development and academic career. When one considers the involvement of other kinds of neuropsychological mechanisms in the writing process (e.g., visual processing, attention, memory, fine motor, organization and planning), the chances of a child experiencing an associated learning problem would appear to increase geometrically. As noted earlier, the occurrence of an isolated writing disorder is relatively rare, and other neuropsychological and academic skill areas always should be assessed. Furthermore, given the convergence of so many neuropsychological abilities in the writing process, a child's failure to transcribe on paper may represent the "tip of the iceberg" with regard to the functioning of that child's central nervous system.

Relatedly, it is clear from the learning disability literature that many children are at risk for poor motivation, frustration, and failure secondary to their learning problems. For example, it could be hypothesized that impaired writing has the potential to result in daily embarrassment or humiliation in school, particularly given that written expression is one of the primary avenues by which children are called on to demonstrate their knowledge. In order to cope with this situation, a student may manifest conduct problems, withdrawn behaviors, or a lack of motivation so as to avoid a writing task or minimize its significance. In addition, primary social-emotional difficulties directly arising from selected kinds of neuropsychological profiles may be manifested (e.g., depression associated with nonverbal learning disabilities; Rourke, 1989). Along with a variety of sociocultural measures, it will be important for the social and emotional functioning to be assessed in children with written expressive language problems.

The cormorbid aspects of written language function and dysfunction are not well understood at this time, particularly with respect to the social -emotional concomitants, and further investigation into this area is warranted. Indeed, it might be that a deeper understanding of the bases for a child's writing problem(s) may assist in developing a treatment program for not only the writing problem but other ongoing concerns as well.

Assessment Techniques and Treatment Linkages

Standardized tests of academic achievement often are insensitive to writing disorders, leaving many children ineligible for individual instruction and remediation efforts (J. Weiner, 1986). Although a number of procedures have been developed to assess handwriting, spelling, and written compositions, there are several assessment factors that must be considered when selecting a test and/or developing an assessment battery for written language expression. In addition to the major assessment domains noted above (i.e., endogenous variables such as neuropsychological functions and personality variables; exogenous variables such as teacher characteristics and emphasis of written language in the child's curriculum; and instrument variables), Luria (1973) noted that assessment should include samples of an individual's copying, writing to dictation, and spontaneous writing (i.e., a sentence, paragraph, or story on a familiar topic). Furthermore, it should be recognized that the nature of the written language prompt (e.g., a picture, starter sentence, or topic) might influence the individual's performance in unknown ways. Examining writing samples directed toward different audiences (e.g., teachers versus peers) and representing different types of writing (e.g., narrative, argumentative) also is encouraged. Furthermore, Reed (1992) noted that assessment parameters should include all of the major production components of a written language task (e.g., spelling, proofreading, mechanics). Relatedly, in diagnosing a written expression disorder, one must attempt to document whether the individual has neuropsychological deficits and whether the writing problem is primary or secondary to oral language or reading disorders, or perhaps secondary to psychiatric problems or sociocultural influences.

Given these general parameters, it is also important to recognize the purposes for performing such a comprehensive assessment. In general, these types of assessments are conducted: 1) to determine the presence of a disorder (i.e., to diagnose) and/or contribute to eligibility for special education services; 2) to gain detailed information on a child's profile of strengths and weaknesses in order to assist in the development of appropriate instructional programs for written language and related classes; 3) to assist in the treatment monitoring process and/or to determine treatment effectiveness; and 4) to provide students with direct feedback on their written language knowledge and productions. Although other questions might be posited, these clearly will be the primary clinical questions that will confront most professionals. At present, there are a variety of indirect and direct assessment methods that have been developed to address these questions.

Indirect Assessment Methods

Indirect measures are so named because they do not require an individual to formulate and produce any original composition. Their primary purpose is to ascertain a student's knowledge about writing conventions as opposed to having the student apply these conventions to an actual writing task (Moran, 1988). In general, these measures tap a student's abilities to detect and correct grammatical errors in sentences, detect and correct spelling mistakes, or, occasionally, rework a sentence or paragraph. Response formats can include cloze procedures, multiple-choice questions, or sentence combining activities.

Indirect measures have the advantage of being relatively objective, with many of these measures being norm referenced (e.g., Woodcock-Johnson Tests of Academic Achievement-Revised [Woodcock & Johnson, 1989)). This advantage increases the likelihood of gaining adequate reliability estimates. However, given that their primary focus is measuring written language conventions (e.g., grammar, punctuation), their use appears to be valid only for questions addressing a child's knowledge of conventions. They appear to be less valid for measuring the composition aspects of writing, although some instruments have attempted to provide estimates of direct as well as indirect aspects of the writing process (e.g., TOWL-2, Woodcock-Johnson Tests of Achievement-Revised). Furthermore, although written language conventions are important to the writing process and, in fact, correlate with general language proficiency (Lloyd-Jones, 1977), they only determine the ability of a student to recognize good writing, not produce it. Consequently, the ecological validity of the indirect measures has been questioned. Their utility for purposes of planning writing instructions appears limited because they cannot determine how well an individual will apply knowledge of the written language conventions (Cooper & Odell, 1977).

Direct Assessment Methods

In contrast to the indirect methods, direct assessment methods require students to produce original discourse. These methods typically are produced following some kind of verbal and/or visual prompt. Good prompts identify the form of the writing to be generated, describe the specific audience to be addressed, and inform the student of the purpose of the communication (Moran, 1988). Direct methods have greater ecological validity with respect to how an individual engages in written discourse, and students have greater control over what they wish to communicate.

A major disadvantage, however, is that such procedures are time consuming to obtain and to score, with few standardized scoring procedures being available. Scoring procedures also may require more intensive training on the part of raters in order to attain acceptable levels of reliability. Typically, direct assessment methods are divided into those designed to assess the processes involved in composing and those designed to assess the product. Although Scardamalia and Bereiter (1986) asserted that the distinction between the two is artificial, largely because they believed the two kinds of writing to be highly interrelated, these methods are reviewed separately here.

Process Methods These kinds of direct assessment strategies are dynamic and involve determining the problem-solving steps or procedures that an individual uses during the written language task (Flower & Hayes, 1980). Procedures for assessing the processes involved in writing can be divided into think-aloud protocols and semi-structured interviews. Think-aloud protocols are the most commonly reported process assessment methods (Emig, 197 1; Flower & Hayes, 1981; Perl, 1979), whereby examiners ask the student to describe his or her thinking during the writing task. The apparent advantage of these types of assessment methods is that they provide glimpses into actual thought processes (e.g., organization, planning, evaluating) while the person is composing. The obvious disadvantage is that the questions may detract from the quality of the written product by interfering with the student's natural flow of thoughts.

Semi-structured interviews are similar to the think-aloud protocols in their purpose, but they typically follow the writing task. Although these interviews are less intrusive on the writing act, they possibly are constrained by memory problems and/or the lack of awareness of a specific process on the part of the writer. Although variants of these procedures have been described (e.g., asking questions only at certain points in the writing process, videotaping), Rose (1984) noted that semi-structured interviews were less accurate than think-aloud protocols in terms of the order of activities and the time allotted to each one.

These types of procedures have provided solid data with respect to the stages of the writing process (Flower & Hayes, 1981). Although clear clinical research methodologies continue to develop in this domain, process-oriented findings provide assistance with the development of treatment programs for children with writing problems.

Product Methods Another type of direct assessment method involves an examination of the writing product. Although work samples can be useful in this regard, there are too many contaminants (e.g., lack of understanding of a task, poor motivation, lack of interest) and, consequently, written products typically are elicited for evaluation purposes. White (1984) noted that reliability of such measures can be satisfactory for a variety of scoring systems, but there are a number of variables (e.g., the writing topic, the nature of the writing scale) that can interfere with the reliability of the scores (Charney, 1984). These types of direct measures, including a variety of curriculum-based assessment strategies (Watkinson & Lee, 1992), elicit original written discourse from students and, thus, tend to have increased content and ecological validity.

Scoring for such evaluation methods is divided into holistic and analytical scoring procedures. Holistic procedures involve global ratings of an entire written language product. Specific rating criteria are defined on a continuum, usually ranging from poor to superior, with interrater agreements being reported as adequate depending on the rigor of training and ongoing compliance with the rating procedures. Despite these findings, Charney (1984) reported that holistic ratings can be influenced by the length of the writing sample; the number of words errors, and unusual vocabulary items; and the neatness of the written discourse. Furthermore, given the unitary score that typically is generated, the utility of such scoring procedures lies solely with diagnostic endeavors and offers little in the way of suggestions for treatment planning.

Analytical or atomistic scoring procedures are more diverse in that they attempt to provide information pertinent to a variety of discrete component skills of the writing product. These procedures have the advantage of examining a written product for strengths as well as weaknesses and, thus, offer direct information to teachers for intervention purposes. Although this latter advantage has been questioned (Parker et al., 1991; White, 1 4), it is clear that t is supposition will be dependent on the number of production components inherent in the analytical scoring procedure selected, and the weight given to each.

Written Expressive Language Measures
Formal Measures These measures tend to be published, standardized measures for indirect and direct aspects of written language assessment. Formal indirect methods, such as many of the written language subtests of the Woodcock-Johnson Tests of Academic Achievement-Revised (Woodcock &Johnson, 1989) focus primarily on written language conventions and less on process or product aspects of a composition. In contrast, formal direct measures tend to focus more on the written language product. A major advantage with the formal measures is that the authors generally have aspired to gain the highest level of psychometric properties possible. However, when one eliminates the formal measures that tap spelling (e.g., Wide Range Achievement Test-Revised [Jastak & Wilkinson, 19841, Test of Written Spelling-2 [Larsen & Hammill, 1986]), and handwriting (e.g., Ayers Handwriting Scale [Ayers, 1920]), there are relatively few instruments available to researchers and clinicians. When tests tapping other aspects of language functioning are eliminated (e.g., Test of Adolescent Language-2 [Hammill, Brown, Larsen, & Wiederholt, 19871, Clinical Evaluation of Language Fundamentals-Revised [Semel, Wiig, & Secord, 19871, Slingerland Screening Tests for Identifying Children with Specific Language Disability [Slingerland, 1970]), there are even fewer that are directly concerned with written language expression. As can be seen in Table 4, there are only a few measures that have been published to date that are exclusively devoted to the assessment of the written language product.

Table 4. Representative formal test batteries for the assessment of written language
Test/author(s) Age range Production components
Picture Story Language Test (Mykelbust, 1965) 7 - 17 Productivity, syntax, and abstraction Test of Written Language-2 (Hammil & Larsen, 1988) 7;6 - 7;11 Vocabulary, spelling, style, logical sentences, sentence combining, thematic maturity, contextual vocabulary, syntactic maturity, contextual style, and three summary indices Test of Early Written Language (Hresko, 1988) 3 - 7 General knowledge of writing purposes, discrimination of visual and nonvisual representational forms, understanding of linguistic terms, and ability to write words, sentences, and stories Writtne Language Assessment (Grill & Kirwin, 1989) >8 General writing ability, productivity, word complexity, readability, and a written language quotient

For example, the Picture Story Language Test (Myklebust, 1965) was perhaps one of the first tasks designed exclusively to measure written expressive language in children ages 7-17. This task consists of a single picture about which the child is requested to compose a story. The writing sample is scored for productivity (e.g., total words, total sentences, words per sentence), syntactic skills, and level of abstract thinking. This test produces age equivalents, percentile ranks, and stanines. Although many of the technical aspects of this test are not satisfactory (e.g., reliability estimates ranging from .34 to .95), this test represents a "landmark test of written language" (Perkins, 1972).

Another formal measure is the TOWL-2 (Hammill & Larsen, 1988). This measure was designed to tap several components (i.e., conventional, linguistic, conceptual) across contrived and spontaneous formats of written language expression. The TOWL-2 was designed for use with children and adolescents ages 7 years, 6 months to 17 years, I I months, with the normative sample being quite large (N = 2,216) and generally representative of most school-age populations. In addition to tapping specific written language conventions, such as spelling, vocabulary, and sentence combining, the structure of the TOWL-2 also requests a spontaneous written language sample generated from a picture stimulus. Psychometric properties of the TOWL-2 appear adequate, although several of the reliability coefficients fall below .80. An alternate form also is available.

Despite the apparent comprehensiveness of this battery, there are a number of concerns that arise with the TOWL-2. There are no strategies recommended for evaluating students who generate unusually short or lengthy compositions (e.g., the use of ratios as opposed to pure numbers), and the all-or-none scoring criteria for several of the subtests do not lend themselves to describing the quality of a written product in an accurate fashion. Nonetheless, this test probably represents one of the most comprehensive formal measures available to researchers and clinicians at the present time.

The Test of Early Written Language (TEWL; Hresko, 1988) represents one of the only formal measures that attempt to address the emergence of written language expression in children ages 3-7. The TEWL employs a variety of items that tap the child's general knowledge of writing purposes, discrimination of verbal and nonverbal visual representational forms, understanding of linguistic terms, and ability to write words, sentences, and stories. The technical characteristics of the TEWL appear nicely developed, with a normative sample of 1,355 children. An interesting feature of the TEWL is that many of the responses are not scored on the basis of exactness, thus allowing the process of written language to unfold for each individual child. Despite these qualities, the ultimate utility of this test remains to be determined.

A final example of a formal measure of written language expression is the Written Language Assessment test (WLA; Grill & Kirwin, 1989). The WLA is a direct, product evaluation instrument that can be used for children ages 8 - 18 years. The manual even states that this test potentially is useful for individuals older than age 18 years, but no data are reported in this regard. An individual's writing is assessed via pictorial and verbal prompts for three writing tasks, each of which represents a different kind of written discourse (i.e., expressive writing, instructive writing, and creative writing). Scores are derived via holistic procedures for General Writing Ability, Productivity, Word Complexity, Readability, and a summative index, the Written Language Quotient. Normative data are available for 1,025 students in grades 3-12. Reliability estimates are variable, ranging from .61 for Readability to .90 for the Written Language Quotient.

Informal Measures In contrast to the small number of formal measures exclusively concerned with written language expression, there are a multitude of informal measures designed to assess the indirect and direct aspects of writing. The primary advantage of these measures is that they can be used easily by a variety of professionals (e.g., teachers) across a variety of settings. Their availability, relative ease of use, and time/ cost-effectiveness also make them prime tools for studying the production components of writing and monitoring a student's progress. A major concern, however, is that their informal nature hinders the development of adequate technical standards. Nonetheless, the contribution of these measures has pushed the importance of establishing assessment treatment linkages in the writing domain, as opposed to instruments designed solely for diagnostic purposes.

For example, Poteet (1980) proposed the Checklist of Written Expression. This checklist tapped several of the major production components of written language, including Penmanship (e.g., spacing on the page, letter formations, pencil grip), Spelling (e.g., letter insertions and omissions, phonetic spelling, letter sequence), Grammar (e.g., capitalization, punctuation, syntax), and Ideation (e.g., type of writing, style, tone). This checklist has significant potential for providing specific information to teachers with respect to an individual student's needs. It also has the potential to provide direct feedback to a student regarding ongoing concerns and progress.

Another informal measure of written language expression has been proposed by E. Weiner (1980). This measure also assesses the direct aspects of written discourse by tapping 41 criteria organized across six major production components: Graphic (visual) Features (e.g., off-line writing); Orthographic (e.g., rule-governed spelling); Phonologic (e.g., sound components); Syntactic (e.g., grammar); Semantic (e.g., vocabulary use); and Self-Monitoring Skills (e.g., revision). Examples of errors or problems are provided for each aspect of the major production components. Weiner has demonstrated the utility of this informal assessment system for the identification, programming, and monitoring of students with written language disabilities.

A third example of an informal measure of written language, the Student Writing Checklist, has been proposed by Isaacson (1988). This checklist includes a variety of subcomponents of written language across six major domains: Ideas/Content (e.g. humor, creativity), Organization (e.g., introduction, clear story line), Vocabulary (e.g., specific nouns and verbs), Sentence Structure (e.g., use of a variety of sentence types), Spelling, and handwriting. All subcomponents are rated via holistic procedures from low to high, with relatively greater weighting being given to Ideas/Content and Organization. The authors note that this checklist should be used across a number of different types of discourse, and that it has the potential to provide a more standardized method for evaluating all kinds of student written productions.

Assessment-Treatment Linkages
Obviously, it is paramount that diagnosticians be clear on the reasons for testing so that specific questions can be addressed. The type of questions being asked, however, also will influence what kinds of instruments and/or scoring procedures will be most useful for the situation at hand. In most instances, teachers and other referral sources wish to know not only diagnostic information but also what strategies would be most useful for treatment and ongoing monitoring

For example, Parker et al. (1991) asserted that writing tests should be sensitive to small increments of skill growth across short and medium periods of time. In fact, these kinds of procedures could be conducted in a cost-efficient manner if used regularly by teachers. Although an initial comprehensive assessment might involve direct and indirect assessment methods, along with process and product evaluation strategies, in order to establish initial goals for intervention, the extensive and highly technical aspects of this kind of assessment likely would not be most efficacious for classroom teachers. More global (e.g., holistic), dynamic assessment strategies might be more useful here. Other assessment findings should be integrated into the total diagnostic picture in planning for the child's writing needs, including specific neuropsychological strengths and weaknesses, comorbid academic problems, and concomitant social-emotional patterns. An assessment of related exogenous variables, such as teacher practices and teacher knowledge of the written language process, also should be included in preparing for treatment planning.

If assessment-treatment linkages for written expressive language are to improve, it will be important for a comprehensive assessment to be accomplished that cuts across all major domains of concern. In addition, continued efforts to clarify the production components involved in the written language process and their (neuro)developmental evolution will be necessary. Furthermore, if there is an underlying multidimensionality to the writing process-and current data would suggest this to be true-an individual's response to specific treatment efforts will be the "acid test" for the clinical utility of any classification scheme. Additional efforts are needed in this regard, and the use of contemporary measurement technology (e.g., growth curve analysis) will be essential to these efforts.

Proposed Directions
Written language expression is a highly complex process that requires interaction between variables intrinsic to the child (e.g., neuropsychological, personality) and variables extrinsic to the child (e.g., teacher-student relationship, teacher knowledge of the writing process). Although investigations of written language expression have been conducted for over 100 years, it has been only recently (i.e., the past 10- 15 years) that a renewed interest has been directed to this area. Part of the reason for this lapse has been ongoing debate as to whether written expression was simply a variant of oral expression; however, in addition, the sheer complexity inherent in studying written language output has presented major hurdles with respect to model development, assessment, and assessment-treatment linkages. Epidemiological data also have been scant, if not nonexistent, although preliminary school-based findings suggest that this area should be of major national concern, particularly given that disrupted writing may represent a variety of other neurodevelopmental concerns (e.g., language deficits, reading problems)-comorbid or otherwise. Furthermore, given emergent information relevant to developmental precursors to the writing process, new models of written expression are likely on the horizon.

With respect to the assessment of written expression, there are a variety of issues that arise. Given the absence of an adequate model for written language, assessment methods have lagged behind those for other academic domains. Similar to the assessment of related academic domains, the assessment of written language expression should be guided by the types of questions being asked. For example, if a referral source is interested in a student's knowledge of grammar, punctuation, and other language conventions, many of the standardized procedures (e.g., Woodcock-Johnson Psychoeducational Battery [Woodcock & Johnson, 19891]) may suffice. However, if the question pertains to the process of generating written discourse, strategies designed to tap these processes should be employed (e.g., think-aloud protocols). Relatedly, if questions relate to a child's ability to generate written discourse and its subsequent quality, direct product-evaluative procedures should be invoked. Although these types of strategies are not earth shattering even to the most inexperienced of professionals, the complex nature of written language offers a variety of routes for the assessment process to travel.

In addition to the relative dearth of instruments exclusively designed to measure written language expression, the currently available tests tend to load heavily on the measurement of conventions. Relatedly, content validity is an overriding problem in all measures of written language because there is not a complete model of written language. This becomes even more complex when integrated with the wide variability of when and how written language is taught in the schools (Salvia & Ysseldyke, 1991). In addition to obtaining estimates of underlying neuropsychological functions, environmental factors, and teacher/school variables, any assessment of writing should include measures of convention as well as a spontaneous writing sample. Although written language conventions are essential to the writing process, direct intervention at this level does not appear to have major impacts on the quality of written discourse. The spontaneous writing sample should be evaluated via a componential analysis of the written product, with perhaps some efforts being devoted to gaining insights into the sequence of events that a child employs during the actual writing process. Whereas global, holistic scoring procedures appear useful for diagnostic purposes, the analytical scoring techniques appear more useful for treatment programming and monitoring. Similarly, by attempting to gain insights into how a child moves through the writing process, specific linkages to interventions are suggested.

For example, if writing is taught as a thinking process containing both product and process goals, the opportunities for children to grow cognitively and affectively are enhanced (Graham, MacArthur, Schwartz., & Page-Voth, 1992). Any model for instruction should encompass the process, product, and purposes of the writing process, and an assessment model should tap these in a corresponding fashion. At present, no single instrument has been designed to address all of the aspects of written language expression, with a noticeable lack of formal instruments being available to clinicians and researchers. Informal instruments appear fruitful, particularly for quick and easy monitoring of intervention procedures, and there are many of them, but they are plagued by a lack of reliability, validity, and general psychometric rigor.

Conclusions
The field of learning disabilities has only begun to address the multitude of issues pertinent to the writing process. Despite these initial efforts, the writing process remains an important area of study, particularly with respect to its relationship to other functional domains in an individual's life. This area is starved for longitudinal study of the development of the writing process, particularly with respect to increasing our knowledge of possible predictors and precursors to the continuum of written language competence. Although several exemplary efforts have been initiated (e.g., Berninger & Hart, in press), additional multivariate efforts to understand written language expression are needed. This information will be useful in the development of a comprehensive model of written language, understanding developmental trajectories and possible syndrome analysis, evolutions of appropriate and comprehensive assessment instruments and procedures, and especially designing individual treatment programs.

References

Abbott, R.D. & Berninger, V.W. (1993). Structural equation modeling of relationships among developmental skills and writing skills in primary- and intermediate-grade writers. Journal of Educational Psychology, 85, 478-508.

American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd ed., rev.). Washington, DC: American Psychiatric Press.

Aram, D., & Nation, J. (1980). Preschool language disorders and subsequent language and academic difficulties. Journal of Child Development, 13, 159-170.

Ayers, L.P. (1920). A Scale for Measuring the Quality of Handwriting of School Children. New York: Russell Sage Foundation.

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Greenwich, CT: JAI Press.

Basso, A., Taborelli, A., & Vignolo, L.A. (1978). Dissociated disorders of speaking and writing in aphasia. Journal of Neurology, Neurosurgery, and Psychiatry, 41, 556-563.

Beauvois, M.G., & D6rouesn6, J. (1981). Lexical or orthographic agraphia. Brain, 104, 21-49.

Beitchman J.H., Nair, R. Ferguson, M., & Patel, P.G. (1986). Prevalence of psychiatric disorders in children with speech and language disorder. Journal of the American Academy of Child Psychiatry, 25, 529-535.

Benson, D.F. (1979). Aphasia, alexia, and agraphia. New York: Churchill Livingston.

Bereiter, C. (1980). Development in writing. In L.W. Gregg & E.R. Steinberg (Eds.). Cognitive processes in writing (pp. 73-93). Hillsdale, NJ: Lawrence Erlbaum Associates.

Berkenkotter, C. (1981), Understanding a writer's awareness of audience. College Composition and Communication, 32, 388-399.

Berninger, V.W., & Fuller, F., (in press). Gender differences in orthographic, verbal, and compositional fluency: Implications for assessing writing disabilities in primary grade children. Journal of School Psychology.

Beminger, V., & Hart, T. (in press). From research to clinical assessment of reading and writing disorders: The unit of analysis problem. In R.M. Joshi & C.K. Leong (Eds.), Reading disabilities: Diagnosis and component processes (pp. 66-89). Kluwer: Academic Publishers.

Berninger, V., Hart, T., Abbott, R.D., & Karovsky, P. (1992). Diagnosing writing disabilities with and without IQ: A flexible, developmental perspective. Learning Disabilities Quarterly, 15, 103-118.

Berninger, V.W., Mizokawa, D.T., & Bragg, R. (1991). Theory-based diagnosis and remediation of writing disabilities. Journal of School Psychology, 29, 57-79.

Berninger, V.W., & Rutberg, J. (1992). Relationship of finger function to beginning writing Application to diagnosis of writing disabilities. Developmental Medicine and Child Neurology, 34, 155-172.

Beminger, V., Yates, C., Cartwright, A., Rutberg, J., Remy, E., & Abbott, R. (1992). Lower-level developmental skills in beginning writing . Reading and Writing: An Interdisciplinary Journal, 4, 257-280.

Botvin, G.J., & Sutton-Smith, B. (1977). The development of structural complexity in children's fantasy narratives. Developmental Psychology, 13, 377-388.

Bracewell, R.J., Frederiksen, C.H., & Frederiksen, J.F. (1982). Cognitive processes in composing and comprehending discourse. Educational Psychologist, 17, 146-164.

Bridge, C.A., & Hiebe'rt, E.H. (1985). A comparison of classroom writing practices, teachers' perceptions of their writing instruction, and textbook recommendations on writing practices. Elementary School Journal, 86, 155-172.

Brophy, J. (1993). Conceptualizing student motivation. Educational Psychologist, 18, 200-215.

Brossell, G., & §ML B.H. (1984). An experiment with the wording of essay topics. College Compmtkom and Communication, 35, 423-425.

Burtis, P., BereiJ& C., Scardamalia, M., & Tetroe, J. (1983). The development of planning in wr In G. Wells & B.

Kroll (Eds.), Explorations in the development of writing (pp. 1 /4). Chichester, England: John Wiley. Cartwright, G ' P. 1969). Written expression and spelling. In R. M. Smith (Ed.), Teacher diagnosis of educational difficulties (pp. 95-117). Columbus, OH: Charles E. Merrill.

Charney, D. (190, . The validity of using holistic scoring to evaluate writing: A critical overview. Research in the Teaching of English, 18, 65-81.

Chedru, F., & GM-MOM, N. (1972). Writing disturbances in acute confusional states. Neuropsychologia, 10, 343-353.

Christenson, S.L. Thurlow, M.L., Ysseldyke, J.E., & McVicar, R. (1989). Written language instruction, for students with mild handicaps: Is there enough quantity to ensure quality? Learning Disabilities Quarterly, 12, 219-229.

Cohn, S.J., & Isaacson, S. (1984). Expository writing skills of academically gifted students (Evaluative Report). Tempe: Arizona State University.

Cooper, C.R., (01"0'). Holistic evaluation of writing. In G.R. Cooper & L. Odell (Eds.), Evaluating writing: Describing, measuring, judging (pp. 3-3 1). Buffalo, NY: National Council of Teachers of English.

Cooper, C.R., & Odell, L. (Eds.). (1977). Evaluating writing: Describing, measuring, judging. Buffalo, NY: National Council of Teachers of English.

Crowhurst, M., Pich6, G. (1979). Audience and mode of discourse: Effects of syntactic complexity in writing on two grade levels. Research in the Teaching of English, 13, 101-109.

D'Angelo, F. (I&DM, Nineteenth-century forms/models of discourse. College Composition and Communication, 35, 31-42.

Educational Testing Service. (1986). Writing objectives: 1988 assessment. National Assessment of Educational Progress. Princeton, NJ: Author.

Ellis, A.W. (19ft. Spelling and writing (and reading and soeaking). In W.W. Ellis (Ed.), Normality and pathology in cognitive functions (pp. 113-146). London: Academic Press.

Ellis, A. (1983). Reading, writing, and dvslexia: A cognitive analysis. Hillsdale, NJ: Lawrence ErIbaum Associates.

Emig, J. (1971), The composing processes of twelfth graders (Research Report No. 13). Urbana, IL: National Council of Teachers of English.

Englert, C.S. (1990). Unraveling the mysteries of writing through strategy instruction. In T.E. Scruggs & B.Y.L. Wong (Eds.), Intervention research in learning disabilities. New York: Springer-Verlag.

Englert, C.S., T. E., Anderson, L.M., Anthony, H.M., Fear, K.L., & Gregg, S. L. (I 988). A case for writing intervention: Strategies for writing informational text. Learning DisR7,71#44, Focus, 3, 98-113.

Englert, C.S., QUM T.E., Anderson, L.M., Gregg, S.L., & Anthony, H.M. (1989). Exposition: Reading, writing, and the metacognitive knowledge of learning disabled students. Learning Disabilities Research, 5, 5-24.

Englert, C.S., & Thomas, C.C. (1987). Sensitivity to text structure in reading and writing: A comparison of learning disabled and non-learning disabled students. Learning Disability Quarterly, 10, 93-105.

Finucci, J.M., Isaacs, S.D., Whitehouse , C.C., & Childs, B. (1982). A quantitative index of reading disability for use in family studies. Developmental Medicine and Child Neurology, 2r, 733-744.

Fitzgerald, J. (1987). Research on revision in writing. Review of Educational Research, 57, 481-506.

Flower, L. (1979). Writer-based prose: A cognitive basis for problems in writing. College English, 41, 19-37.

Flower, L.S., & Hayes, J.R. (1980). The dynamics of composing: Making plans and juggling constraints. In L.W. Gregg & E.R. Steinbert (Eds.), Cognitive processes in writing (pp. 31-50). Hillsdale, NJ: Lawrence Erlbaum Associates.

Flower, L., & Hayes, J. (1981). A cognitive process theory of writing. College Composition and Communication, 32, 365-387.

Frith, U. (1983). The similarities and differences between reading and spelling problems. In M. Rutter (Ed.), Developmental neuropsychiatry (pp. 453-472). New York: Guilford Press.

Gaddes, W.J. (1985). Learning disabilities and brain function: A neuropsychological approach (2nd ed.). New York: Springer-Verlag.

Gagne, R.M. (1985). The conditions of learning. A theory of instruction (4th ed.). New York: Hold, Rinehart, & Winston.

Goldstein, K. (1948). Language and language disturbance. New York: Grune & Stratton. Graham, S., & Harris, K. R., (1987). Improving composition skills of inefficient learners with self- instructional strategy training. Topics in Language Disorders, 7, 66-77.

Graham, S., & Harris, K.R. (1989). A components analysis of cognitive strategy training: Effects on learning disabled students' compositions and self-efficacy. Journal of Educational Psychology, 81, 353-361.

Graham, S., Harris, K. R., & MacArthur, C. A. (1990). Learning disabled and normally achieving students' knowledge of the writing process. Unpublished raw data.

Graham, S., Harris, K.R., MacArthur, C., & Schwartz, S. (1991). Writing and writing instruction for students with learning disabilities: Review of a research program. Learning Disabilities Quarterly, 14, 89-114.

Graham, S., MacArthur, C., Schwartz, S., & Pacye-Voth, V. (1992). Improving the compositions of students with learning disabilities using a strategy involving product and process goal setting. Exceptional Children, 58, 322-326.

Greenberg, K. ( 1981 ). The effects of variations in essay questions on the writing performance of CUNY freshman. New York: The City University of New York Instructional Resource Center.

Gregg, N. (1991). Disorders of written expression. In A. Bain, L.L. Bailet, & L.C. Moats (Eds.), Written language disorders: Theory into practice (pp. 65-97). Austin, TX: PRO-ED.

Gregg, N. (1992). Expressive writing. In S.R. Hooper, G.W. Hynd, & R.E. Mattison (Eds.), Developmental disorders: Diagnostic criteria and clinical assessment (pp. 127172). Hillsdale, NJ: Lawrence Erlbaum Associates.

Gregg, N., & McAlexander, P.A. (1989). The relation between sense of audience and specific learning disabilities: An exploration. Annals of Dyslexia, 39, 206-226.

Grill, J.J., & Kirwin, M.M. (1989). Written language assessment. Novato, CA: Academic Therapy Publications.

Halliday, M.A.K., & Hasan, R. (1976). Cohesion in English. London: Longman Group,

Hammill, D., Brown, Larsen, S., & Wiederholt, L. (1987). Test of Adolescent Language-2. Austin, TX: PRO-ED.

Hammill, D., & Larsen, S. (1988). Test of Written Language-Revised. Austin, TX: PROED.

Hayes, J. (1990). Individuals and environments in writing instruction. In B.F. Jones & L. Idol (Eds.), Dimensions of thinking and cognitive instruction (pp. 241-263). Hillsdale, NJ: Lawrence Erlbaum Associates.

Hayes, J.R., & Flower, L.S. (1980). Identifying the organization of writing processes. In L.W. Gregg & E. R. Steinbert (Eds.), Cognitive processes in writing (pp. 3-30). Hillsdale, NJ: Lawrence Erlbaum Associates.

Hayes, J. R., & Flower, L. S. 1986). Writing research and the writer. American Psychologist, 4 1, 1106-1113. Hayes, J., Flower, L.S., Schriver, P., Stratemen, D., & Carey, W. (1989). Cognitive processes in revision (Tech. Rep. No. 12). Pittsburgh, PA: Carnegie Melon University, Communication Design Center.

Head, H. (1926). Aphasia and kindred disorders of speech. Cambridge: Cambridge University Press.

Hdcaen, H., & Albert, M.L. (1978). Human neuropsychology. New York: John Wiley & Sons.

Hoetker, J., & Brossell, G. ( 1989). The effects of systematic variations in essay topics on the writing performance of college freshman. College Composition and Communication, 40, 414-421.

Hooper, S.R., & Willis, W.G. (1989). Learning disability subtyping:Neuropsychological foundations, conceptual models, and issues in clinical differentiation. New York: Springer-Verlag.

Hotopf, N. (1980). Slips of the pen. In U. Frith (Ed.), Cognitive processes in spelling. London: Academic Press.

Howie, S.H. (1984). A guide for teaching writing in content areas. Boston: Allyn & Bacon. Hresko, W.P. (1988). Test of Early Written Language. Austin, TX: PRO-ED. Interagency Committee on Learning Disabilities. (1987). Learning disabilities: A report to the U.S. Congress. Washington, DC: Author.

Isaacson, S. (1988). Effective instruction in written language. In E.L. Meyen, G.A. Vergason, & R.J. Whelan (Eds.), Effective instructional strategies for exceptional children (pp. 288-306). Denver, CO: Love Publishing Company.

Jastak, J., & Wilkinson, G. (1984). Wide Range Achievement Test-Revised. Wilmington, DE: Jastak Associates.

Johnson, D., & Myklebust, H. ( 1967). Learning disabilities: Educational principles and practices. New York: Grune & Stratton.

Kaplan, E., & Goodglass, H. (1981). Aphasia-related disorders. In M.T. Sarno (Ed.), Acquired aphasia (pp. 256-272). New York: Academic Press.

Kavale, K. (1987). Theoretical issues surrounding severe discrepancy. Learning Disabilities Research, 3, 12-20.

King, M.L., & Rentel, V.M. (1981). Research update: Conveying meaning in written texts. Language Arts, 58, 721-728.

Larsen, S.C., & Hammill, D.D. (1986). Test of Written Spelling-2. Austin, TX: PROED.

Leinhardt, G., Zigmond, N., & Cooley, W.W. (1980, April). Reading instruction and its effects. Paper presented at the annual meeting of the American Educational Research Association, Boston.

Lemer, J.W. (1976). Children with learning disabilities: Theories, diagnosis, teaching strategies. Boston, MA: Houghton Mifflin.

Lichtheim, L. (1885). On aphasia. Brain, 7, 433-484.

Litowitz, B. (1981). Developmental issues in written language. Topics in Language Disorders, 1, 73-89.

Lloyd-Jones, R. (1977). Primary trait scoring. In C. Cooper & L. Odell (Eds.), Evaluating writing (pp. 33-68). Urbana, IL: National Council of Teacher of English.

Luria, A.R. (1970). The functional organisation of the brain. Scientific American, 222, 66-78.

Luria, A.R. (1973). The working brain: An introduction to neuropsychology. New York: Basic Books.

Luria, A.R. (1980). Higher cortical functions in man. New York: Basic Books,

Lyon, G.R. (1987). Learning disabilities research: False starts and broken promises. In

S. Vaughn & C. Bos (Eds.), Research in learning disabilities: Issues and future directions (pp. 69-85). San Diego, CA: College Hill Press.

Lyon, R., & Moats, L. (1989). Critical issues in the instruction of the learning disabled. Journal of Consulting and Clinical Psychology, 56, 830-855.

MacArthur, C., & Graham, S. (1987). Learning disabled students' composing with three methods: Handwriting, dictation, and word processing. Journal of Special Education, 21, 22-42.

MacArthur, C., & Schwartz, S. (in press). An integrated approach to writing instruction: The computer in writing instruction project. Learning Disability Forum.

Maxwell, S., & Wallach, G. (1984). The language-learning disabilities connection: Symptoms of early language disability change over time. In G.P. Wallach & K.G. Butler (Eds.), Language learning disabilities in school-age children (pp. 15-34). Baltimore: Williams & Wilkins.

McCutchen, D., & Perfetti, C. (1982). Coherence and connectedness in the development of discourse production. Text, 2, 113-119.

Moffett, J. (1986). Teaching the universe of discourse. Boston: Houghton Mifflin. Montague, M., Graves, A., & Leavell, A. (1991). Planning, procedural facilitation, and narrative composition of junior high students with learning disabilities. Learning Disabilities Research and Practice, 6, 219-224.

Moran, M.R. (1988). Options for written language assessment. In E.L. Meyen, G.A. Vergason, & R.J. Whelan (Eds.), Effective instructional strategies for exceptional children (pp. 465-480). Denver, CO: Love Publishing Company.

Myklebust, H. (1965). Development and disorders of written language: Vol. 1. New York: Grune & Stratton.

Myklebust, H. (1973). Development and disorders of written language: Vol. 2. Studies of normal and exceptional children. New York: Grune & Stratton.

National Assessment of Educational Progress. (1980a). Writing achievement, 1969-79: Results from the third national writing assessment: Vol. 1. 17-year-olds (Report NO. I I O-W-0 1). Denver, CO: Author.

National Assessment of Educational Progress. (1980b). Writing achievement, 1969-79: Results from the third national writing assessment: Vol. 2. 13-year-olds (Report No. I I O-W-02). Denver, CO: Author.

National Assessment of Educational Progress. (1980c). Writing achievement, 1969-79: Results from the third national writing assessment: Vol. 3. 9-year-olds (Report No. I 10W-03). Denver, CO: Author.

National Assessment of Educational Progress. (1990). National Center for Education Statistics, Office of Educational Research and Improvement, U.S. Department of Education. Washington, DC: U.S. Government Printing Office.

Nelson, H.E., & Warrington, E.K. (1976). Developmental spelling retardation. In R.M. Knights &

D.J. Bakker (Eds.), The neuropsychology of learning disorders (pp. 325332). Baltimore: University Park Press.

Newcomer, P.L., & Barenbaum, E.M. (1991). The written composing ability of children with learning disabilities: A review of the literature from 1980 to 1990. Journal of Learning Disabilities, 24, 578-593.

Newcomer, P., Nodine, B., & Barenbaum, E. (1988). Teaching writing to exceptional children: Reaction and recommendations. Exceptional Children, 54, 559-561.

Nielson, J.M. (1946). Agnosia, apraxia, aphasia: Their value in cerebral localization (2nd ed.). New York: Hoeber.

Ogle, J.W. (1867). Aphasia and agraphia. Report of the Medical Research Counsel of Saint George's Hospital, 2, 83-121

O'Hare, A.E., & Brown, J.K. (1989). Childhood dysgraphia, Part 1: An illustrated clinical classification. Child Care, Health, and Development, 15, 79-104.

Parker, R.I., Tindal, G., & Hasbrouck, J. (1991). Progress monitoring with objective measures of writing performance for students with mild disabilities. Exceptional Children, 58, 61-73.

Perkins, W.H. (1972). Review of the Picture Story Language Test. In O.K. Buros (Ed.), Seventh mental measurements yearbook (Vol. 1, pp. 206-207). Highland Park, NJ: Gryphon Press.

Perl, S. (1979). The composing processes of unskilled college writers. Research in the Teaching of English, 13, 317-336.

Phelps-Gunn, T., & Phelps-Terasaki. D. (1982). Written language instruction: Theory and remediation. Rockville, MD: Aspen Systems.

Poteet, J,A. (1980). Informal assessment of written expression. Learning Disabilities Quarterly, 3, 88-98.

Reed, M.S. (1992). Educational assessment. In M.D. Levine, W.B. Carey, & A.C. Crocker (Eds.), Developmental-behavioral pediatrics (2nd ed., pp. 638-645). Philadelphia: W.B. Saunders Company.

Resnick, L.B. (1985). Cognition and instruction: Recent theories of human competence. In B.L. Hammonds (Ed.), Psychology and learning (pp. 123-186). Washington, DC: American Psychological Association.

Roeltgen, D. (1985). Agraphia. In K. M. Heilman & E. Valenstein (Eds.), Clinical neuropsychology (pp. 75-96). New York: Oxford University Press.

Roeltgen, D., & Heilman, K.M. (1984). Lexical agraphia. Brain, 107, 811-827. Roit, M.L., & McKenzie, R.G. (1985). Disorders of written communication: An instructional priority for LD students. Journal of Learning Disabilities, 19, 258-260.

Rose, M. (1984). Writer's block: The cognitive dimension. Carbondale, IL: Southern Illinois University Press.

Rosenthal, T.L., & Zimmerman, B.J. (1978). Social learning and cognition. New York: Academic Press.

Rourke, B. P. (1989). Nonverbal learning disabilities: The syndrome and the model. New York: Guilford Press.

Rourke, B.P. (1991). Neuropsychological validation of learning disability subtypes. New York: Guilford Press.

Rubin, D. (1982). Adapting syntax in writing to varying audiences as a function of age and social cognitive ability. Journal of Child Language, 9, 497-5 10.

Salvia, J., & Ysseldyke, J.E. (1991). Assessment (5th ed.). Boston: Houghton Mifflin.

Sandler, A.D., Watson, T.E. Footo, M., Levine, M.D., Coleman, W.L.. & Hoper, S.R. (1992). Neurodevelopmental study of writing disorders in middle childhood. Developmental and Behavioral Pediatrics, 13, 17-23.

Scardamalia, M., & Bereiter, C. (1986). Research on written composition. In M. C. Wittrock (Ed.), Handbook on research on teaching (3rd ed., pp. 778-803). New York: Macmillan.

Schenck, S.J. (1981). The diagnostic/ instructional link in individualized education programs. Journal of Special Education, 14, 337-345.

Semel, E., Wiig, E.H., & Secord, W. (1987). Clinical evaluation of language fundamentals-Revised. San Antonio, TX: The Psychological Corporation.

Shallice, T. (1988). From neuropsychology to mental structure. New York: Cambridge University Press.

Shaughnessy, M.P. (1977). Errors and expectations: A guide for the teacher of basic writing. New York: Oxford University Press.

Siegel, L. (1989). IQ is irrelevant to the definition of learning disabilities. Journal of Learning Disabilities, 22, 469-486.

Slingerland, B. ( 1970). Teacher's manual to accompany Slingerland Screening Tests for Identifying Children with Specific Language Disability. Los Angeles, CA: Western Psychological Services.

Smith, W.L., Hull, G.A., Land, R.E.. Moore, M.T., Ball, C., Dunham, D.E., Hickey. L.S., & Rizich, C.W. (1985). Some effects of varying the structure of a topic of college students' writing. Written Communication, 2, 73-89.

Smith, W.L., & Swan, M.B. (1978). Adjusting syntactic structures to varied levels of audience. Journal of Experimental Education, 46, 29-34.

Sommers, N. (1980). Revision strategies of student writers and experienced adult writers. College Composition and Communication, 31, 378-388.

Stanovich, K.E. (1986). Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Research Quarterly, 21, 360-407.

Stanovich, K.E. (1992). Developmental reading disorder. In S.R. Hooper, G.W. Hynd, & R.E. Mattison (Eds.), Developmental disorders: Diagnostic criteria and clinical assessment (pp. 173-208). Hillsdale, NJ: Lawrence Erlbaum Associates.

Stark, R. Tallal, P., & Mellits, D. (1983). Behavioral attributes of language disorders in children. In C. Ludlow (Ed.), Genetic aspects of speech and language disorders (pp. 37-51). New York: Academic Press.

Stein, N. L., & Trabasso, T. (1982). What's in a story: An approach to comprehension and instruction. In R. Glaser (Ed.), Advances in instructional psychology (pp. 213-267). Hillsdale, NJ: Lawrence Erlbaum Associates.

Straw, S.B. (1981). Grammar and teaching of writing: Analysis versus synthesis. In V. Froese & S.B. Straw (Eds.), Research in the language arts: Language and schooling (pp. 147-161). Baltimore: University Park Press,

Sweeney, J.E., & Rourke, B.P. (1978). Neuropsychological significance of phonetically accurate and phonetically inaccurate spelling errors in younger and older retarded spellers. Brain and Language, 6, 212-225.

Temple, C. M. (1988). Developmental dyslexia and dysgraphia persistence in middle age. Journal of Communication Disorders, 21, 189-207.

Tindal, G., & Parker, R. (199 1). Identifying measures for evaluating written expression. Learning Disabilities Research and Practice, 6, 211-218.

U.S. Office of Education, (1976). Public Law 94-142 regulations: Proposed rulemaking.

Federal Register, 41, 52404-52407. Vygotsky, L. S. (1962). Thought and language. Cambridge, MA: MIT Press.

Watkinson, J.T., & Lee, S.W. (1992). Curriculum-based measures of written expression for learning-disabled and nondisabled students. Psychology in the Schools, 29, 184 - 191.

Weiner, E. (1980). The diagnostic evaluation of writing skills (DEWS). Learning Disabilities Quarterly, 3, 54-59.

Weiner, J. (1986). Alternatives in the assessment of the learning disabled adolescent: A learning strategies approach. Learning Disabilities Focus, 1, 97-107.

White, E. (1984). Holisticism. College Composition and Communication, 35, 400-409.

Wing, A.M., & Baddeley, A.D. (1980). Spelling errors in handwriting: A corpus and a distributional analysis. In U. Frith (Ed.), Cognitive processes in spelling. London: Academic Press.

Wong, B.Y.L. (1991). Three conceptual perspectives on the connections between reading and writing processes. In A. McKeough & J. L. Lupart (Eds.), Toward the practice of theory-based instruction (pp. 66-91). Hillsdale, NJ: Lawrence Erlbaum Associates.

Wong, B., Wong, R., & Blankenship, J. (1989). Cognitive and metacognitive aspects of learning disabled adolescents' composition problems. Learning Disabilities Quarterly, 15, 145-152.

Woodcock, R.W., & Johnson, M.B. (1989). Woodcock-Johnson Psychoeducational Battery-Revised. Allen, TX: DLM.

Zimmerman, B.J. (1990). Social cognitive theory and self-regulated learning. In B.J. Zimmerman & D.H. Schunk (Eds.), pp. 1-25). New York: Springer-Verlag.