Writing Interventions for High School Students With Disabilities

Abstract

Proficiency in writing increases opportunities in higher education, employment, and social relationships. Many students in the United States, however, are struggling writers. In particular, the writing performance of students with disabilities in secondary grades continues to lag behind grade level peers. The purpose of this article was to review writing interventions for high school students with disabilities to identify those practices that have been shown to be effective in supporting writing development. All studies were peer-reviewed, employed a single-case experimental design, and evaluated writing interventions used with students with disabilities in Grades 9 through 12. We used the What Works Clearinghouse (WWC) single-case design standards (Kratochwill et al., 2010) to evaluate the effects of writing interventions in the 14 identified studies. Specific interventions included Self-Regulated Strategy Development (SRSD), Expressive Writing™ (Direct Instruction), Cognitive Strategy Instruction in Writing, and pentop computers. Results indicated that only two SRSD programs and the pentop computer program demonstrated strong or moderate effects on written expression in studies that met WWC evidence standards with reservations.

Keywords

writing secondary students student with disabilities writing interventions high school students single-case design research

With the world’s ever increasing dependency on the Internet and other forms of electronic communication, the ability to communicate via writing is an invaluable ticket to the future; writing proficiency will open opportunities to higher education, employment, social relationships, and better quality of life (Graham & Perin, 2007a; Kiuhara, Graham, & Hawken, 2009; National Commission on Writing, 2008; Rogers & Graham, 2008; Taft & Mason, 2011). Nevertheless, large numbers of students in the United States struggle with writing, and many in this group are students with disabilities (National Center for Education Statistics, 2007). Furthermore, since the implementation of No Child Left Behind in 2001, writing receives less emphasis in the classroom than reading and math (National Commission on Writing, 2006).

Insufficient attention to writing instruction often persists through high school. Even though most high school students do not possess the writing skills that will be required in higher education or postsecondary employment, 39% of high school seniors reported rarely writing an extensive report or paper longer than two pages (National Commission on Writing, 2008). Almost half of high school teachers responding to a national survey admitted that they did not assign complex writing tasks even once a month, and many students are not making the progress needed to be successful in college or the workforce (Kiuhara et al., 2009; National Commission on Writing, 2008). In addition, the majority of teens surveyed by the National Commission on Writing (2008) acknowledged that more writing instruction in school would improve their writing skills. In the Kiuhara et al. (2009) report on a national survey of writing practices among high school teachers, recommendations included the need for (a) increased use of evidence-based practices in writing instruction, (b) more frequent adaptation of writing instruction to meet the needs of struggling writers, and (c) better teacher training in writing instruction. Unfortunately, the results of the 2011 National Assessment of Educational Progress in writing indicated that students with disabilities continue to lag behind their grade level peers: Only 5% of students with disabilities as compared with 27% of overall students in Grades 8 and 12 scored at or above a proficient level (National Center for Education Statistics, 2012; Shah, 2012).

In recent years, the amount of research conducted on interventions for struggling writers has increased (Graham & Perin, 2007a). However, the majority of research on writing interventions for students with disabilities has been conducted on students with learning disabilities (LD) and has tended to focus on late elementary through middle grades (Baker, Chard, Ketterlin-Geller, Apichatabutra, & Doabler, 2009; Joseph & Konrad, 2009). Studies on writing interventions for adolescents and, in particular, high school students have been sparse (Rogers & Graham, 2008).

Previous Reviews on Writing Interventions

Since 2001, 10 reviews have been published on writing instruction or interventions that included students with disabilities. Most reviews highlighted interventions with students with LD or at risk for LD (Baker et al., 2009; Gersten & Baker, 2001; Graham & Harris, 2003; Graham & Perin, 2007a, 2007b; Mason & Graham, 2008), but there were also reviews on students with intellectual disabilities (ID; Graham & Harris, 2003; Joseph & Konrad, 2009), students with autism spectrum disorders (ASD; Pennington & Delano, 2012), students with disabilities other than LD (Taft & Mason, 2011), and struggling writers (Rogers & Graham, 2008). The various reviews evaluated single-case design research, group experimental or quasi-experimental designs, or some combination of these research designs. Some reviews targeted a specific disability, age group, or research design, while others focused on writing programs, such as Self-Regulated Strategy Development (SRSD). Grade levels included in the studies spanned from kindergarten through Grade 12.

Authors conducting these reviews used a variety of tools to analyze their selected studies. In most of the reviews, in addition to synthesizing study components for readers, reviewers of group experimental and quasi-experimental research calculated effect sizes of writing intervention outcomes using Cohen’s d, and reviewers of single-case design used percent nonoverlapping data points (PND) to estimate intervention effect. In more recent years, reviewers have begun to address the level of quality of the research study itself in addition to reporting effect sizes of intervention outcomes. In a 2005 issue of Exceptional Children, researchers in special education proposed the use of quality indicators in the planning and evaluating of special education research: Gersten et al. (2005) presented quality indicators for group experimental and quasi-experimental special education research, while Horner et al. (2005) recommended quality indicators for single-case design research. Three of the more current reviews on writing interventions included quality indicator scores using Gersten et al. standards (Baker et al., 2009; Graham & Perin, 2007a; Mason & Graham, 2008), and three other recent reviews included Horner et al.’s quality indicators (Baker et al., 2009; Mason & Graham, 2008; Rogers & Graham, 2008). As Odom et al. (2005) stated in the same issue of Exceptional Children, quality indicators “serve as guidelines for . . . reviewers who evaluate the ‘believability’ of research findings” (p. 141).

Results of a majority of the reviews indicated that, of all writing interventions studied, strategy instruction and SRSD programs were consistently found to be effective and evidence-based in both group design and single-case design research (Baker et al., 2009; Graham & Harris, 2003; Graham & Perin, 2007b; Rogers & Graham, 2008), with elementary and middle school students (Graham & Harris, 2003), and for students with ID (Joseph & Konrad, 2009), LD (Baker et al., 2009), and disabilities other than LD (Taft & Mason, 2011). In addition to SRSD and strategy instruction, other writing interventions were shown to be effective practices, including use of a word processor (Graham & Perin, 2007b; Rogers & Graham, 2008), grammar instruction, prewriting activities (Rogers & Graham, 2008), reading summarization strategies, and peer assistance and collaboration (Graham & Perin, 2007a). Furthermore, Gersten and Baker (2001) found that all interventions designed to improve expressive writing for students with LD had positive effects on writing quality.

It is important to note that, although many high-quality reviews on writing interventions have been published over the past dozen years and across all grade levels, none have been conducted specifically on writing interventions for high school students with disabilities. In addition, no previous reviews on writing interventions have evaluated single-case design research according to the new standards proposed by What Works Clearinghouse (WWC; Kratochwill et al., 2010). The new WWC standards recommend initial steps for determining whether a single-case research design meets standards, and if minimum standards are evident, subsequent steps for determining whether the intervention produced an effect on the outcome variables. Thus, the current review was undertaken as an avenue to analyze writing interventions from two novel perspectives: first, from a focus specifically on writing interventions conducted with high school students with disabilities to determine the most effective methods with this particular population, and second, as an opportunity to evaluate research using the recently released WWC single-case design standards (Kratochwill et al., 2010). In sum, this review evaluated single-case design research on writing interventions for high school students with disabilities. We posed two research questions:

Research Question 1: In all studies meeting criteria for the review, which writing interventions were used and which disabilities were targeted?

Research Question 2: Which writing intervention studies for high school students with disabilities met WWC evidence standards for single-case design research with or without reservations and demonstrated strong or moderate effect of a causal relation between the intervention and the outcome variables?

Method

The following criteria were used for identifying articles on writing interventions for high school students with disabilities: (a) research must have been peer-reviewed, published between 1965 and 2011, and employed single-case designs capable of demonstrating experimental control (i.e., reversal, multiple baseline, or alternating treatment; Horner et al., 2005); (b) the focus of the interventions was on improving written expression (e.g., planning, composing, editing, revising) rather than spelling, grammar, or other elements of writing mechanics; (c) at least some participants in the study were identified as having a disability; (d) participants in the study were identified as being in high school, 9th through 12th grades, or at least 14 years old if the grade level was not specified; and (e) the article was printed in English. In addition to targeting single-case designs, this review followed the direction of several other reviews in focusing on explicit instruction in written expression (Gersten & Baker, 2001; Graham & Perin, 2007a, 2007b; Joseph & Konrad, 2009; Mason & Graham, 2008; Taft & Mason, 2011). Interventions on handwriting and spelling were not included in the current review even though research has demonstrated some impact of these interventions on writing development (Graham & Harris, 2009).

The initial method for locating research studies consisted of an online search of the following databases: Academic Search Complete, Education Research Complete, Eric, Medline, and PsychINFO. Key words used in the search were adopted from two meta-analyses on writing interventions (Graham & Perin, 2007a; Rogers & Graham, 2008). The terms writing, composition, and narrative were paired with dictation, genre, genre and instruction, goal setting, inquire, models, peer collaboration, peer revising, peers, planning, revising, pre-writing, process writing, reinforcement, self-evaluation, self-monitoring, sentence combining, strategy instruction, summary instruction, summary strategies, summary writing, technology, word processing, word processor, writer’s workshop, and writing assessment.

Results from databases were combined to yield a total of 136 articles. Abstracts were investigated and potential articles were printed for further examination. Articles not meeting criteria for this review were discarded. The bulk of discards fell into three groups: research articles on interventions in areas other than writing (n = 37), research design was not single-case (n = 28), or students in the studies did not meet criteria for age range or grade level (n = 36). Other discarded articles included reviews or reports rather than a research study (n = 15), students without disabilities (n = 3), and studies with handwriting or spelling interventions (n = 9). At this point, we had 8 studies meeting criteria for review.

Next, the tables of content and abstracts of the following journals were reviewed either online or manually for a period extending from 1965 through November 2011: Behavioral Disorders, Educational Psychologist, Exceptional Children, Focus on Autism and Other Developmental Disabilities, Journal of Applied Behavior Analysis, Journal of Autism and Developmental Disorders, Journal of Behavioral Education, Journal of Educational Research, Journal of Learning Disabilities, Journal of Special Education Technology, Learning Disability Quarterly, Learning Disabilities Research and Practice, and Remedial and Special Education. One relevant study was uncovered from journal searches (n = 1). Then, an ancestral search was conducted using reference lists of all selected studies and several relevant writing intervention reviews (Baker et al., 2009; Graham & Harris, 2003; Graham & Perin, 2007a; Joseph & Konrad, 2009; Mason & Graham, 2008; Rogers & Graham, 2008; Taft & Mason, 2011), leading to the discovery of additional studies (n = 2). Leading researchers in the area of writing interventions for learners with disabilities were also contacted via email to locate overlooked articles and uncover the latest research in this area (M. Delano, S. DeLaPaz, S. Graham, K. Harris, M. Konrad, R. Kubina, C. MacArthur, L. Mason, and R. Reid). One article available online and 2 articles accepted but not yet published were located in this manner (n = 3). Finally, an additional ancestral search was completed with newly acquired articles to ensure that all possible relevant articles were found (n = 0). In the end, a total of 14 studies, shown in Table 1, were considered eligible for this review.

Table 1.

Overview of Reviewed Single-Case Design Studies on High School Writing Interventions.

Authors	Disability (age)	Design	Intervention and genre	Graphed measures	Maintenance/generalization	Procedural fidelity/social validity
Bouck, Doughty, Flanagan, Szwed, and Bassette (2010)	ID (n = 2) LD (n = 1) (15–18)	Multiple probe across part.	Pentop computer Compare–contrast and persuasive essays	Quality (Likert-type scale)	N/Y	Y/Y
Delano (2007)	AS (n = 2) (15–17)	Multiple baseline across responses	SRSD via video self-modeling Persuasive essays	Word count Functional elements	Y/Y	N/N
Dowell, Storey, and Gleason (1994)	LD (n = 3) (14–16)	Multiple baseline across part.	Comparison of process method, process method with visuals, and systematic instruction Descriptive paragraphs	Quality (Likert-type scale) Percent descriptive words	N/N	Y/Y
Hoover, Kubina, and Mason (2012)	LD (n = 4) (16–19)	Multiple probe across part.	SRSD 10-min persuasive quick writes	Response parts Word count	Y/N	Y/Y
Jacobson and Reid (2010)	ADHD (n = 3)	Multiple probe across part.	SRSD Persuasive essays	Quality (Likert-type scale) Word count Essay elements	Y/N	Y/N
Kiuhara, Graham, and Hawken (2009)	ADHD (n = 2) LD (n = 2) ED (n = 1) SLI (n = 1) (15–16)	Multiple probe across part.	SRSD Persuasive essays	Quality (primary trait scale) Essential elements Functional elements	Y/N	Y/Y
Konrad, Trela, and Test (2006)	ID (n = 4) (15–18)	Multiple baseline across part.	SRSD Paragraphs on IEP goals	Quality (paragraph scoring guide) Content	Y/Y	Y/Y
Liberty and Fitzpatrick (1994)	ID (n = 6) (14–18)	Multi-treatment ABCBC	Teacher modeling and demonstrating Narrative writing	Lexical ties	N/N	N/N
MacArthur and Philippakos (2010)	LD (n = 3) (14)	Multiple probe across part.	SRSD Compare–contrast essays	Quality (primary trait rubric) Text structure elements	Y/N	Y/Y
Mason, Kubina, and Hoover (2011)	ED (n = 3) (15–17)	Multiple baseline across part.	SRSD 10-min persuasive quick writes	Quality (holistic measure) Response parts Word count	Y/N	Y/Y
Miller and Lignugaris-Kraft (2002)	LD (n = 2) (14–17)	Multiple probe across text structures	CSIW Paragraph responses to test questions	t units	Y/Y	Y/Y
Viel-Ruma, Houchins, Jolivette, Fredrick, and Gama (2010)	LD (n = 3) LD/ELL (n = 3) (14–17)	Multiple probe across part.	Expressive Writing (DI), abbreviated version 3-min narrative samples	CWS Word count	Y/N	Y/Y
B. Walker, Shippen, Alberto, Houchins, and Cihak (2005)	LD (n = 3) (15–16)	Multiple probe across part.	Expressive Writing (DI) 3-min narrative samples	CWS	Y/Y	Y/Y
B. D. Walker, Shippen, Houchins, and Cihak (2007)	LD (n = 3) (15)	Multiple probe across part.	Expressive Writing (DI) 3-min narrative samples	CWS	Y/Y	Y/Y

Note. ID = intellectual disabilities; LD = learning disabilities; part. = participants; N = no; Y = yes; AS = Asperger syndrome; SRSD = Self-Regulated Strategy Development; ADHD = attention-deficit/hyperactivity disorder; ED = emotional and/or behavioral disorder; SLI = speech-language impairments; IEP = Individualized Education Program; CSIW = Cognitive Strategy Instruction in Writing; ELL = English language learner; DI = Direct Instruction; CWS = correct word sequences.

Coding of Study Features

After the articles were selected, the first and second author read and coded the articles according to WWC standards (Kratochwill et al., 2010). To collect interobserver agreement (IOA) data on coding, the two authors conducted coding independently from each other. The first step was to evaluate the research designs using Kratochwill et al. (2010) design characteristics to verify that designs were single-case (e.g., ABAB design, multiple baseline designs): (a) the intervention was applied to and the outcome was measured on individual cases; (b) the individual case served as “its own control for purposes of comparison” (p. 14), which in most single-case designs is indicated by baseline conditions; and (c) the dependent variables were measured repeatedly and across conditions. The second step was to code for evidence of WWC design standards: (a) the researcher demonstrated systematic manipulation of the independent variable (i.e., intervention); (b) outcome variables were measured systematically, with IOA data collected on at least 20% of the data points in each condition and with an IOA percentage agreement of 0.80 or higher (or with Cohen’s kappa of .60 or higher); (c) at least three attempts to demonstrate experimental control were presented at three different points in time; and (d) a minimum of three data points per condition to Meet Standards With Reservations, or a minimum of five data points per condition to Meet Standards. For the first two steps, studies were coded dichotomously (i.e., yes, no) on each standard. If any standards in the first or second steps were coded “no,” with the exception of the last standard (minimum of five data points per condition), the study was rated as Does Not Meet Standards and was eliminated from further evaluation. Studies meeting standards, with or without reservations, were moved into the third step of coding: visual analysis of graphed data.

During visual analysis, baseline and intervention mean levels were calculated, and trends (“the slope of the best-fitting straight line for data within a phase”; Kratochwill et al., 2010, p. 5) and variability (“as reflected by the data’s range or standard deviation”; Kratochwill et al., 2010, p. 5) were evaluated. Kratochwill et al. (2010) did not specify methods for determining variability; therefore, steps provided by Gast and Spriggs (2010) were used to determine whether the data for each condition would be considered stable or variable (using a 20% stability envelope, stability was indicated by 80% of data points falling within 20% of the median level). Finally, the percent of overlapping data points across adjacent conditions was calculated. At this point, each graph was coded dichotomously for whether (a) the baseline condition demonstrated a predictable pattern of responding and a need for intervention, (b) the intervention produced immediacy of effect (i.e., an immediate change in the level was shown between the last three data points in one condition and the first three data points of the next), and (c) similar conditions demonstrated similar patterns of responding (e.g., data paths of all baseline and intervention conditions for a dependent variable were compared; Kratochwill et al., 2010). Each graph depicting a single participant or case and showing a predictable baseline with a contra-therapeutic level and trend, followed by an immediate change in the level with stable data in the intervention condition, represented one demonstration of effect; conversely, any failure to demonstrate experimental control (e.g., no change in level or trend between conditions, highly variable data, unpredictable patterns, or delayed response to treatment) was considered a demonstration of noneffect. The fourth and final step of the coding process involved an overall integration of visual analysis across cases in a study to determine whether experimental effect of the intervention was demonstrated: Strong Evidence of a Causal Relation was determined by at least three demonstrations of effect with no noneffects, Moderate Evidence of a Causal Relation was determined by at least three demonstrations of effect with at least one demonstration of noneffect, and No Evidence of a Causal Relation was indicated with less than three demonstrations of effect (Kratochwill et al., 2010).

Reliability measures were calculated on the first and second author’s agreement in coding WWC standards. Level, trend, variability, predictable patterns, immediacy, and similar patterns were evaluated for each participant (or case) within each condition and across adjacent conditions of each study. A point-by-point method was used for calculating IOA, in which the two authors totaled the number of items that they scored identically, then divided that number by the total number of items assessed, and finally, multiplied by 100 to reach the percent of agreement (Ayres & Gast, 2010). IOA for the first step (i.e., verification of single-case design) and second step (i.e., verification of evidence of WWC design standards, including meeting WWC minimum thresholds for IOA, demonstrations of effect, and data points per condition), leading to the determination of whether the studies met research design standards, was 100%. IOA for the third step (i.e., visual analysis of baseline and intervention conditions for level, trend, variability, and predictable data patterns within and across conditions; immediacy of effect and overlapping data points in adjacent conditions; and similar patterns in similar conditions) was 98% (range = 96%–100%). IOA for the fourth step, overall evidence of experimental effect, was 100%.

Results

Participants and Settings

A total of 51 high school students with disabilities participated in the 14 studies reviewed. As described in Table 1, the participants—31 males and 20 females—were 14 to 19 years old. Participants were identified with LD (n = 27), ID or cognitive disabilities (n = 12), attention deficit hyperactivity disorders (ADHD; n = 5), emotional or behavioral disorders (EBD; n = 4), Asperger syndrome (AS; n = 2), or speech-language impairments (n = 1). Ethnic information was not provided in all studies, but the largest group represented was Caucasian (n = 16), and other ethnic groups included African American (n = 7) and Hispanic or Latino (n = 6). Most studies were published since 2001 and included programs of intervention such as SRSD (n = 7), Direct Instruction (n = 3), and Cognitive Strategy Instruction in Writing (CSIW; n = 1), as well as one study that evaluated the use of pentop computers for prewriting and drafting prompts. The remaining studies were published in 1994, evaluating interventions in process approach versus systematic instruction (n = 1) and models and demonstrations (n = 1).

The single-case design studies in this review were conducted within the United States with the exception of one New Zealand study (Liberty & Fitzpatrick, 1994). Ten studies reported geographical locations within the United States, including four Southeastern, two Eastern, two Western, and two Midwestern locations. Five studies took place in urban areas and four in suburban areas. All except two studies were conducted in public schools; Delano (2007) worked with students in a private office setting, and MacArthur and Philippakos (2010) worked with students in a private school for students with disabilities.

Independent Variables—Writing Interventions

SRSD

SRSD teaches explicit strategies that lead students to self-regulation in planning, writing, and editing; the level of teacher support decreases as the student demonstrates understanding of the writing process (Graham & Harris, 2009). Writing instruction is individualized and students must meet the criteria before progressing to a new level of instruction (Graham & Harris, 2003). Graham and Harris (2009) listed the six stages of SRSD: (a) Develop and Activate Background Knowledge, (b) Discuss It, (c) Model It, (d) Memorize It, (e) Support It, and (f) Independent Performance. In SRSD studies, mnemonics are employed to help participants remember steps of writing strategies.

Persuasive writing skills were the focus of the majority of SRSD studies at the high school level, with two studies focusing on persuasive essays (Jacobson & Reid, 2010; Kiuhara, O’Neill, Hawken, & Graham, 2012) and two additional studies targeting 10-min persuasive quick writes (Hoover, Kubina, & Mason, 2012; Mason, Kubina, & Hoover, 2011). In addition, Delano (2007) conducted a study that combined SRSD with video modeling to boost persuasive essay writing skills. Of the two remaining SRSD studies, one focused on writing compare–contrast essays (MacArthur & Philippakos, 2010) and the other addressed writing paragraphs in which students proposed goals that they wanted to include in their Individualized Education Programs (IEPs; Konrad, Trela, & Test, 2006).

Direct Instruction and other interventions

Five additional studies were published over the past decade, including three studies using a Direct Instruction program. Expressive Writing™ (Direct Instruction), Level 1 (Engelmann & Silbert, 1983), was used to deliver instruction on mechanics, sentence writing, paragraph and story writing, and editing (Viel-Ruma, Houchins, Jolivette, Fredrick, & Gama, 2010; B. Walker, Shippen, Alberto, Houchins, & Cihak, 2005; B. D. Walker, Shippen, Houchins, & Cihak, 2007). Viel-Ruma et al. (2010) extended B. Walker et al.’s (2005) study by implementing an abbreviated Expressive Writing™ program, teaching students odd-numbered lessons only.

Interestingly, only one computer-based study was conducted at the high school level; students were provided prewriting, planning, and drafting prompts for persuasive and compare–contrast essays with pentop computers (Bouck, Doughty, Flanagan, Szwed, & Bassette, 2010). In another study, Miller and Lignugaris-Kraft (2002) used a cognitive strategy, CSIW, to teach text structures in attempt to improve students’ paragraph-writing ability in academic test responses. CSIW, bearing similarities to SRSD, teaches students to apply strategies related to planning, organizing, writing, editing, and revising; self-regulation is encouraged (Miller & Lignugaris-Kraft, 2002).

Of studies published prior to the millennium, only two met criteria for this review, both published in 1994. Dowell, Storey, and Gleason (1994) initially compared the effectiveness of a general writing process approach with a writing process with visual prompts, and then to a systematic, direct instruction program. Liberty and Fitzpatrick (1994) compared the effects of teacher modeling versus demonstrating how to write a narrative.

Dependent Variables—Targeted Skills

Writing quality

Writing quality was not consistently defined or measured across studies. Seven studies in all measured writing quality using a variety of Likert-type scoring procedures, rubrics, or scoring guides. In two studies, quality was evaluated by rating the appropriate use of essay writing components, such as planning, organization, content, style, voice, and grammar (Bouck et al., 2010; MacArthur & Philippakos, 2010). Three studies developed scales using anchor papers of peers to determine quality (Dowell et al., 1994; Jacobson & Reid, 2010; Mason et al., 2011). In another study, Konrad et al. (2006) awarded quality points for including a topic sentence, using supporting details, presenting information logically, using transition words, restating the topic in a concluding sentence, and avoiding irrelevant information. A final study measuring quality used a rubric to measure persuasiveness, awarding a low score for not taking a persuasive position and a high score for taking a strong position supported by examples, refuting counterarguments, and displaying a well-organized flow of thoughts (Kiuhara et al., 2012).

Essay elements

Seven studies measured the number of essay elements, also termed response parts, essay parts, text structure elements, or essential elements. Of the seven, five studies evaluated the number of persuasive essay parts such as topic sentence, supporting ideas, rejection of opposing arguments, and concluding sentence (Delano, 2007; Hoover et al., 2012; Jacobson & Reid, 2010; Kiuhara et al., 2012; Mason et al., 2011). In a study on compare–contrast writing, MacArthur and Philippakos (2010) awarded text structure points for introduction, hook, comparison, topic sentence, details, conclusion, summary, and transitions. In addition, Miller and Lignugaris-Kraft (2002) used t-unit measures to determine percentage of correct text structure elements containing relevant content on paragraphs written in response to reading-passage questions.

Number of words

Five studies measured the number of words written. Three of the five studies used a word processor to count words in transcribed papers (Hoover et al., 2012; Jacobson & Reid, 2010; Mason et al., 2011). Hoover et al. (2012) and Mason et al. (2011) confirmed word processor counts with manual tallies; Jacobson and Reid (2010) employed a second transcriber for 20% of student papers. The remaining two studies simply used a manual count of words (Delano, 2007; Viel-Ruma et al., 2010).

Other dependent variables

Other dependent variables included correct word sequences (CWS), paragraph content, percent of descriptive words, and number of lexical ties. Three studies measured CWS (Viel-Ruma et al., 2010; B. Walker et al., 2005; B. D. Walker et al., 2007), which are curriculum-based measures that positively correlate with writing quality (Videen, Deno, & Marston, 1982); every two words in a sequence are examined together to determine the number of correct word pairings (Viel-Ruma et al., 2010). Another study evaluated paragraph content using a scoring guide to rate items as being not evident, a partial response, or a complete response (Konrad et al., 2006). Dowell et al. (1994) measured the percent of descriptive words used. Finally, Liberty and Fitzpatrick (1994) counted the number of complex lexical ties used in student writing. Liberty and Fitzpatrick defined lexical ties as use of complex semantic relationships such as word repetition, personal pronouns, demonstratives, comparatives, and conjunctions.

Implementers

Special education teachers implemented the interventions in three studies (Bouck et al., 2010; Dowell et al., 1994; Viel-Ruma et al., 2010). In addition, a researcher who was also a teacher implemented interventions in three studies (Hoover et al., 2012; Liberty & Fitzpatrick, 1994; Mason et al., 2011). Researchers implemented interventions in the remaining studies.

Procedural Fidelity

Procedural fidelity assessments were reported for 12 studies, with 11 studies assessing reliability of at least 20% of sessions. Four studies had observers use checklists to review audiotaped or videotaped instruction (Hoover et al., 2012; Kiuhara et al., 2012; MacArthur & Philippakos, 2010; Mason et al., 2011), and 5 additional studies used direct observation with task analysis steps or lesson plan checklists (Bouck et al., 2010; Jacobson & Reid, 2010; Konrad et al., 2006; Miller & Lignugaris-Kraft, 2002; Viel-Ruma et al., 2010). Two other studies used observation of Direct Instruction procedures to measure percentages of student responses, teacher praise, and teacher correction procedures (B. Walker et al., 2005; B. D. Walker et al., 2007). Finally, Dowell et al. (1994) reported that two audiotaped sessions were reviewed to determine that the teacher was following scripted lessons. Two studies did not report procedural fidelity measures (Delano, 2007; Liberty & Fitzpatrick, 1994).

Social Validity

Eleven studies reported social validity measures. Measures consisted of surveys or interviews with participants, teachers, and parents. These social validity measures suggested a general approval of the writing interventions. In addition, 2 studies used peer comparisons to indicate social validity of the outcome (Dowell et al., 1994; Miller & Lignugaris-Kraft, 2002). Both studies reported that participants’ written products at the end of the studies were rated as comparable with papers written by some general education peers. Three studies did not report social validity measures (Delano, 2007; Jacobson & Reid, 2010; Liberty & Fitzpatrick, 1994).

WWC Standards

Evaluation of single-case design

Kratochwill et al. (2010) described types of single-case research that could potentially exhibit three demonstrations of effect, including withdrawal (e.g., ABAB) and multiple baseline designs. Of the 14 studies, 13 used multiple baseline or multiple probe designs, with 11 employing multiple baseline or multiple probe across participants, presenting the intervention to participants across different points in time. Two other studies implemented multiple baseline across behaviors, applying the intervention to different writing elements (Delano, 2007) or text structures (Miller & Lignugaris-Kraft, 2002) at different points in time for each participant. Liberty and Fitzpatrick (1994) used a multitreatment design (ABCBC), similar to a withdrawal design, to compare the effects of two different interventions.

To be recognized as Meets Evidence Standards (with or without reservations) according to WWC standards, the intervention must be “systematically manipulated” by the researcher, data must be collected over time with IOA data collected in a minimum of 20% of sessions in each condition, meeting minimal agreement levels (e.g., 0.80 percentage agreement, .60 Cohen’s kappa), and there must be at least “three demonstrations of experimental effect at three different points in time with a single case” (Kratochwill et al., 2010, pp. 14–15). All 14 studies provided descriptions of the baseline condition, the intervention, and the process for beginning intervention, so all studies were considered to provide evidence of manipulation of the independent variable. In addition, all 14 studies reported using IOA to verify dependent measures, but only 7 met IOA criteria for evidence standards. More specifically, 8 of the studies reported collecting IOA in 20% of each condition, but only 7 studies reported IOA scores at or above the minimum agreement levels.

After evaluating all requirements for WWC single-case design standards, none of the 14 studies were considered to Meet Evidence Standards for Design, but 6 Met Evidence Standards With Reservations. Only one study met the WWC design criteria of five or more data points per condition (Hoover et al., 2012). However, that study failed to meet minimum thresholds for IOA and, therefore, did not meet the WWC design standards. In addition, the 6 studies that met evidence standards with reservations, as shown in Table 2, met all WWC design standards with the exception of having only three or four data points per condition instead of five. Each of the 6 studies meeting standards (with reservations) presented graphed data on 1 to 3 outcome variables, resulting in a total of 12 outcome variables being available for further evaluation to consider the strength of evidence for a causal relation.

Table 2.

Application of What Works Clearinghouse (WWC) Criteria to Determine Which Single-Case Designs Meet Evidence Standards.

Study	Independent variable manipulated	Reported IOA data collection in at least 20% of conditions & participants	IOA scores at or above 80%	Three attempts to demonstrate effect	Minimum of three data points per phase	Minimum of five data points per phase	WWC design standards results
Bouck, Doughty, Flanagan, Szwed, and Bassette (2010)	Y	Y	Y	Y	Y	N	MWR
Delano (2007)	Y	N	Y	N	Y	N	Does not meet
Dowell, Storey, and Gleason (1994)	Y	Y	Y	Y	Y	N	MWR
Hoover, Kubina, and Mason (2012)	Y	N	N	Y	Y	Y	Does not meet
Jacobson and Reid (2010)	Y	Y	Y	Y	N	N	Does not meet
Kiuhara, Graham, and Hawken (2009)	Y	Y	Y	Y	Y	N	MWR
Konrad, Trela, and Test (2006)	Y	N	N	Y	Y	N	Does not meet
Liberty and Fitzpatrick (1994)	Y	N	N	Y	N	N	Does not meet
MacArthur and Philippakos (2010)	Y	Y	Y	Y	Y	N	MWR
Mason, Kubina, and Hoover (2011)	Y	Y	Y	Y	Y	N	MWR
Miller and Lignugaris-Kraft (2002)	Y	Y	Y	N	N	N	Does not meet
Viel-Ruma, Houchins, Jolivette, Fredrick, and Gama (2010)	Y	N	Y	Y	Y	N	Does not meet
B. Walker, Shippen, Alberto, Houchins, and Cihak (2005)	Y	N	Y	Y	Y	N	Does not meet
B. D. Walker, Shippen, Houchins, and Cihak (2007)	Y	Y	Y	Y	Y	N	MWR
Percentages	100%	57%	79%	86%	79%	7%	43% (MWR)

Note. 1OA = interobserver agreement; Y = yes; N = no; MWR = meets with reservations.

Strong evidence of effect

In the MacArthur and Philippakos (2010) study, an SRSD program was implemented with compare–contrast writing. One outcome variable, text structure elements, provided three clear demonstrations of effect at three points in time. Even though there were some variable data points in the baseline condition, the predicted baseline pattern indicated continuation of low levels. In addition, the data points after SRSD instruction demonstrated immediate strong level increases for all participants.

Moderate evidence of effect

Two studies presented evidence of moderate effect (i.e., three or more demonstrations of effect with at least one demonstration of noneffect) with one outcome variable each. In one study, Bouck et al. (2010) measured the writing quality of persuasive and compare–contrast essays with the use of a pentop computer; visual analysis indicated an immediate increase in essay quality for all participants when the pentops were used. One participant (Cara) discontinued use of the pentop computer toward the end of the intervention condition, and her quality scores dropped. Moderate evidence of effect was demonstrated because of decelerating trend during the latter part of the Cara’s intervention condition. In another study, Kiuhara et al. (2012) measured the effects of teaching SRSD for persuasive writing; when measuring the number of essential writing elements used in an essay, all participants demonstrated increases in level and accelerating trend in the intervention condition, as well as sufficient immediacy of effect. However, the two final baseline data points of one graph (Brianna’s) were on a slope that continued through the first data points of the intervention condition; therefore, this graph was considered a demonstration of noneffect.

No evidence of effect

Five studies provided no evidence of effect for at least one outcome variable. Dowell et al. (1994) presented highly variable data for two outcomes—use of descriptive words and writing quality—and although means (averages) of data within conditions showed increases from baseline to the first intervention condition (process method with visuals) and again to the second intervention condition (direct systematic instruction), functional control was not evident. In addition, even though Kiuhara et al. (2012) demonstrated moderate evidence of effect for one outcome variable, no effect was evident for two other outcome variables: use of functional elements (one graph showed a trend in the intervention condition that was not unlike baseline) and holistic quality (data points showed no immediacy of effect in the first tier of participants, and the intervention data for all participants were variable). In the MacArthur and Philippakos (2010) study, even though another outcome variable demonstrated strong evidence of effect, the measure of overall quality showed no evidence of effect because participants in the third tier had highly variable data in intervention condition. The Mason et al. (2011) study was unable to demonstrate the effect for outcome measures of quality, number of response parts, or number of words; data were highly variable in the intervention condition for quality and number of response parts (number of words was not analyzed for effect because IOA data were not reported). Finally, B. D. Walker et al. (2007) measured CWS to determine the effects of Expressive Writing™ instruction; however, the lack of response by the third participant at the beginning of the intervention condition prevented three demonstrations of effect.

Discussion

As expected from the previous reviews of the literature, half of the studies in the current review evaluated the effects of SRSD as an intervention, and students with LD tended to be the target population. In addition, studies in this review reported that SRSD was used for writing instruction with students with ADHD, emotional and/or behavioral disorder (ED), ID, and speech-language impairments. Another study with students with LD evaluated a cognitive strategies program (CSIW) that had some elements in common with SRSD. Other writing interventions were analyzed as well, including three studies investigating a Direct Instruction program (Expressive Writing™) with students with LD. Surprisingly, only one study evaluated the use of technology, providing pentop computers to students with ID and LD. Older studies evaluated the effects of different writing instruction methods with students with ID and LD.

The current review’s application of WWC standards for single-case design studies involved a different evaluation process from that of previous reviews, which used PND calculations or Horner et al.’s (2005) quality indicators to evaluate the effect. Using WWC standards, only 6 of the 14 single-case design studies met design standards with reservations, and of those, only 2 of 7 SRSD studies and the 1 pentop computer study demonstrated sufficient evidence of experimental control with an outcome variable (i.e., Strong or Moderate Evidence of Causal Relations). In contrast, previous reviews that used PND to determine effectiveness found most strategy instruction interventions, including SRSD, SRSD with video modeling, and CSIW, to be effective with several outcome variables (Graham & Perin, 2007b; Joseph & Konrad, 2009; Taft & Mason, 2011). Likewise, 1 review that used Horner et al.’s quality indicators found strategy instruction to be effective (Rogers & Graham, 2008), and in 2 other reviews, SRSD was determined to have a sufficient research base to meet the Horner et al. criteria for evidence-based practices (Baker et al., 2009; Mason & Graham, 2008). Of special interest were results from direct comparison of the studies evaluated by both the current and previous reviews. For example, 3 SRSD studies that were reported as effective in previous reviews failed to meet evidence standards for single-case designs under WWC standards—Konrad et al. (2006), reviewed by Joseph and Konrad (2009); Delano (2007), reviewed by Pennington and Delano (2012) as well as Taft and Mason (2011); and Jacobson and Reid (2010), also reviewed by Taft and Mason.

The differences in the conclusions of the current review and prior reviews suggest that WWC standards may be more stringent than Horner et al.’s (2005) quality indicators. In fact, WWC standards specify minimum thresholds for IOA and data points per condition, whereas the Horner et al. quality indicators do not. Under WWC single-case design evaluation, studies are not examined for treatment effects if minimum thresholds of IOA are not reported (i.e., 20% of data points in every condition of the study). In contrast, although Horner et al. suggested similar minimum thresholds of IOA and advised that IOA be collected in each condition, the proportion to be collected per condition was not stipulated. In addition, even though Horner et al. recommended that baseline conditions should have at least five data points and that the dependent variable should be measured repeatedly during the intervention conditions, WWC went further by prescribing a minimum number of data points per condition to meet design standards (three data points to meet with reservations, five data points to fully meet the design standards). However, although the two sets of single-case research design standards exhibit some differences, many elements of the two approaches are similar. For example, both employ visual analysis to determine functional control and the effectiveness of interventions using single-case design methodology, and both define functional control as three demonstrations of experimental effect at three different points in time (Horner et al., 2005; Kratochwill et al., 2010).

Although WWC standards (Kratochwill et al., 2010) are executed differently and are more procedurally specific than Horner et al.’s (2005) quality indicators, we are in agreement with Wolery (2013) that the WWC standards are not and should not be the definitive tool for determining effective interventions in single-case design research. One reason for taking this position is that WWC standards did not include all of the guidelines suggested by Horner et al. For instance, WWC guidelines did not address the importance of procedural fidelity or social validity measures, even though both are listed as quality indicators in Horner et al. In addition, as Wolery highlighted, it is important for researchers in single-case design to provide detailed descriptions of all experimental conditions, as well as of the participants and the research setting, to add to the replicability (and external validity) of the research. While Horner et al. addressed the need for providing detailed descriptions in research reporting, WWC standards did not. Perhaps a combination of the two sets of guidelines—WWC standards and Horner et al. quality indicators—would produce the most comprehensive method of evaluating the quality of research and determining the evidence of intervention effect.

Limitations

The current review had several limitations. Primarily, because selection criteria included only single-case design articles from peer-reviewed publications on high school students with disabilities, relevant information on other writing interventions was likely precluded from this study. More specifically, research from group comparison studies, non-peer-reviewed research (e.g., in book chapters, doctoral dissertations), and studies involving younger students or students without disabilities were not considered in this review. Another limitation, simply attributed to the nature of single-case designs, was that the number of total participants in this review comprised a small sample size. Finally, although both authors received basic training in graduate level single-case design courses, neither held certification in visual analysis, a WWC requirement for determining Strong Evidence of a Causal Relation between an intervention and an outcome (Kratochwill et al., 2010).

Implications for Future Research

To meet WWC design standards, future researchers need to plan a priori to collect sufficient amount of data, with the ideal goal of five or more data points per condition. Only one of the studies in the current review met this criterion. In addition, more attention should be given not only to the collection of IOA on dependent measures but also to clear reporting of IOA results. Eight studies (57%) reported collecting IOA in 20% of all conditions; other studies reported IOA but failed to clearly indicate the conditions in which IOA data were collected. Ultimately, six studies (43%) met design standards with reservations, and of these, only three studies provided demonstrations of causal effect on outcome variables. The two main problems in demonstrating experimental control were (a) the intervention was introduced before the baseline data were stable and predictable and (b) the intervention condition had variable data. Planning a research study with WWC standards in mind will only serve to strengthen the study.

Furthermore, more single-case design research is warranted for writing interventions for students of all disability areas at the high school level to determine the most effective and efficient practices that will also promote long-term maintenance and generalization of writing skills. Single-case design research should be expanded to include other writing interventions that have been targeted with younger students or with group experimental or quasi-experimental studies. In this era of technology, it was surprising to not find more single-case design studies targeting computer programs or applications that would promote writing proficiency for high school students. MacArthur and Philippakos (2010) observed that more research is warranted for high school students with disabilities to improve writing skills in academic content areas. In addition, more information is needed on the effects of teachers implementing writing interventions in their own classrooms, including the general education setting. Furthermore, for true comparisons in determining the most efficient interventions, researchers should consider collecting data during the instructional process rather than waiting until instruction is completed. Finally, it was notable that no practice met the minimum WWC requirements that would allow studies to be combined for an overall summary rating. To justify a summary rating, Kratochwill et al. (2010) indicated that the treatment must be investigated by at least five single-case studies that meet minimum evidence standards (with at least three different research teams, three different locations, and a minimum of 20 participants or single-case examples). Hence, more research studies are needed to replicate findings highlighted in this review to recommend the practices with confidence.

Implications for Practice

Because no practice met the minimum WWC requirements for an overall summary rating, caution should be used when implementing writing interventions with high school students with disabilities; teachers should monitor student outcomes with formative assessments to assess individual progress. In addition, according to the results of the current review, which applied WWC single-case design standards, only two SRSD programs and the pentop computer technology could be tentatively recommended as positively affecting writing outcomes for high school students with disabilities (i.e., demonstrating strong or moderate evidence of effect). Specifically, two studies that met WWC evidence standards with reservations found SRSD to have strong and moderate effects, respectively, for compare–contrast and persuasive essay writing. Using the steps in SRSD to teach students strategies for planning and writing compare–contrast essays, MacArthur and Philippakos (2010) showed that students increased their use of text structures, such as introduction, comparisons, details, summary, and conclusion. Similarly, Kiuhara et al. (2012) used SRSD to teach planning, drafting, and writing strategies. Kiuhara et al. found that, with SRSD instruction, students increased the number of essential elements used in writing persuasive essays, including providing the context for the problem, a problem statement, support for the position taken, and a concluding statement. An especially salient observation from both SRSD studies was that skills were maintained (at 2 to 3 weeks for Kiuhara et al., 2012, and 4 to 8 weeks for MacArthur & Philippakos, 2010) after the intervention supports were removed.

The other intervention shown to have moderate effects by a study that met WWC standards with reservations was the pentop computer, which resulted in an improvement in the overall quality of both compare–contrast and persuasive essays (Bouck et al., 2010). The pentop computer provided step-by-step vocal directions and prompts to guide students in developing an organizational map and a draft outline of either a compare–contrast or persuasive essay; students wrote the final version of their essays independently, without the pentop computer. Quality ratings on students’ planning, organization, content, style, and grammar showed immediate improvements when the pentop computer was used to plan and organize essays. Participants did not maintain quality in their essay writing after discontinuing the use of the pentop computer—in contrast to the two SRSD interventions, which showed maintenance of writing skills after the intervention ended. Overall, the implications from this review are that SRSD holds the most promise as an effective writing intervention that will also promote maintenance of writing skills for compare–contrast and persuasive essays.

In conclusion, many high school students with disabilities struggle with writing to express thoughts and communicate ideas. Results of this review indicate that, even though a small number of writing practices have demonstrated tentative effectiveness, there is a continued need for more systematic and well-executed research to guide educators in best practices of writing instruction and intervention. WWC single-case design standards, while not the only tool available, can be invaluable for planning and implementing a study, as well as for evaluating the quality of a research design and the effectiveness of a completed study. A well-designed study can add to the research base and contribute to the acceptability of an intervention as an evidence-based practice—a practice that has the potential to improve a student’s quality of life, the ultimate goal of educational research.

Footnotes

Acknowledgements

The authors wish to acknowledge the support and advisement of Dr. Kristin Sayeski in preparation of this manuscript.

Declaration of Conflict of Interest

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Ayres

Gast

D. L.

(2010). Dependent measures and measurement procedures. In Gast

D. L.

(Ed.), Single subject research methodology in behavioral sciences (pp. 129–165). New York, NY: Routledge.

Baker

S. K.

Chard

D. J.

Ketterlin-Geller

L. R.

Apichatabutra

Doabler

(2009). Teaching writing to at-risk students: The quality of evidence for self-regulated strategy development. Exceptional Children, 75, 303–318.

Bouck

E. C.

Doughty

T. T.

Flanagan

S. M.

Szwed

Bassette

(2010). Is the pen mightier? Using pentop computers to improve secondary students’ writing. Journal of Special Education Technology, 25(4), 33–47.

Delano

M. E.

(2007). Improving written language performance of adolescents with Asperger Syndrome. Journal of Applied Behavior Analysis, 40, 345–351.

Dowell

H. A.

Storey

Gleason

(1994). A comparison of programs designed to improve the descriptive writing of students labeled learning disabled. Developmental Disabilities Bulletin, 22, 73–91.

Engelmann

Silbert

(1983). Expressive writing I. Chicago, IL: SRS/McGraw-Hill.

Gast

D. L.

Spriggs

A. D.

(2010). Visual analysis of graphic data. In Gast

D. L.

(Ed.), Single subject research methodology in behavioral sciences (pp. 199–233). New York, NY: Routledge.

Gersten

Baker

(2001). Teaching expressive writing to students with learning disabilities: A meta-analysis. The Elementary School Journal, 101, 251–272.

Gersten

Fuchs

L. S.

Compton

Coyne

Greenwood

Innocenti

M. S.

(2005). Quality indicators for group experimental and quasi-experimental research in special education. Exceptional Children, 71, 149–164.

10.

Graham

Harris

K. R.

(2003). Students with learning disabilities and the process of writing: A meta-analysis of SSRD studies. In Swanson

Harris

Graham

(Eds.), Handbook of learning disabilities (pp. 323–344). New York, NY: Guilford Press.

11.

Graham

Harris

K. R.

(2009). Almost 30 years of writing research: Making sense of it all with The Wrath of Khan. Learning Disabilities Research & Practice, 24, 58–68.

12.

Graham

Perin

(2007a). A meta-analysis of writing instruction for adolescent students. Journal of Educational Psychology, 99, 445–476.

13.

Graham

Perin

(2007b). What we know, what we still need to know: Teaching adolescents to write. Scientific Studies of Reading, 11, 313–335. doi:10.1080/10888430701530664

14.

Hoover

T. M.

Kubina

R. M.

Jr. Mason

L. H.

(2012). Effects of Self-Regulated Strategy Development for POW+TREE on high school students with learning disabilities. Exceptionality, 20, 20–38. doi:10.1080/09362835.2012.640903

15.

Horner

R. H.

Carr

E. G.

Halle

McGee

Odom

Wolery

(2005). The use of single-subject research to identify evidence-based practice in special education. Exceptional Children, 71, 165–179.

16.

Jacobson

L. T.

Reid

(2010). Improving the persuasive essay writing of high school students with ADHD. Exceptional Children, 76, 156–174.

17.

Joseph

L. M.

Konrad

(2009). Teaching students with intellectual or developmental disabilities to write: A review of the literature. Research in Developmental Disabilities, 30, 1–19. doi:10.1016/j.ridd.2008.01.001

18.

Kiuhara

S. A.

Graham

Hawken

L. S.

(2009). Teaching writing to high school students: A national survey. Journal of Educational Psychology, 101, 136–160.

19.

Kiuhara

S. A.

O’Neill

R. E.

Hawken

L. S.

Graham

(2012). The effectiveness of teaching 10th grade students STOP, AIMS, and DARE for planning and drafting persuasive text. Exceptional Children, 78, 335–355.

20.

Konrad

Trela

Test

D. W.

(2006). Using IEP goals and objectives to teach paragraph writing to high school students with physical and cognitive disabilities. Education and Training in Developmental Disabilities, 41, 111–124.

21.

Kratochwill

T. R.

Hitchcock

Horner

R. H.

Levin

J. R.

Odom

S. L.

Rindskopf

D. M.

Shadish

W. R.

(2010). Single-case design technical documentation. Retrieved from http://ies.ed.gov/ncee/wwc/pdf/reference_resources/wwc_scd.pdf

22.

Liberty

K. A.

Fitzpatrick

(1994). Effects of models and demonstrations of personal narratives on the use of lexical ties by writers with mental retardation. Journal of Behavioral Education, 4, 7–20.

23.

MacArthur

C. A.

Philippakos

(2010). Instruction in a strategy for compare-contrast writing. Exceptional Children, 76, 438–456.

24.

Mason

L. H.

Graham

(2008). Writing instruction for adolescents with learning disabilities: Programs of intervention research. Learning Disabilities Research & Practice, 23, 103–112. doi:10.1111/j.1540-5826.2008.00268.x

25.

Mason

L. H.

Kubina

R. M.

Jr. Hoover

(2011). Effects of quick writing instruction for high school students with emotional disturbances. Journal of Emotional and Behavioral Disorders. Advances online publication. doi:10.1177/1063426611410429

26.

Miller

T. L.

Lignugaris-Kraft

(2002). The effects of text structure discrimination training on the writing performance of students with learning disabilities. Journal of Behavioral Education, 11, 203–230.

27.

National Center for Education Statistics. (2007). The nation’s report card: Writing 2007. Washington, DC: Institute of Education Sciences, U.S. Department of Education. Retrieved from http://nces.ed.gov/nationsreportcard/pdf/main2007/2008468.pdf

28.

National Center for Education Statistics. (2012). The nation’s report card: Writing 2011 (NCES 2012-470). Washington, DC: Institute of Education Sciences, U.S. Department of Education.

29.

National Commission on Writing. (2006). Writing and school reform. Retrieved from http://www.collegeboard.com/prod_downloads/writingcom/writing-school-reform-natl-comm-writing.pdf

30.

National Commission on Writing. (2008). Writing, technology and teens. Retrieved from http://www.collegeboard.com/prod_downloads/prof/community/PIP_Writing_Report_FINAL.pdf

31.

Odom

S. L.

Brantlinger

Gersten

Horner

R. H.

Thompson

Harris

K. R.

(2005). Research in special education: Scientific methods and evidence-based practices. Exceptional Children, 71, 137–148.

32.

Pennington

R. C.

Delano

M. E.

(2012). Writing instruction for students with autism spectrum disorders: A review of the literature. Focus on Autism and Other Developmental Disabilities, 27, 158–167.

33.

Rogers

L. A.

Graham

(2008). A meta-analysis of single subject design writing intervention research. Journal of Educational Psychology, 100, 879–906.

34.

Shah

(2012, September 18). Fraction of special education students proficient on national writing test. Education Week. Retrieved from http://blogs.edweek.org/edweek/speced/2012/09/fraction_of_special_ed_students.html

35.

Taft

R. J.

Mason

L. H.

(2011). Examining effects of writing interventions: Highlighting results for students with primary disabilities other than learning disabilities. Remedial and Special Education, 32, 359–370.

36.

Videen

Deno

S. L.

Marston

(1982). Correct word sequences: A valid indicator of proficiency in written expression (Research Report No. 84). Minneapolis: University of Minnesota, Institute for Research on Learning Disabilities.

37.

Viel-Ruma

Houchins

D. E.

Jolivette

Fredrick

L. D.

Gama

(2010). Direct instruction in written expression: The effects on English speakers and English language learners with disabilities. Learning Disabilities Research & Practice, 25, 97–108. doi:10.1111/j.1540-5826.2010.00307.x

38.

Walker

Shippen

M. E.

Alberto

Houchins

D. E.

Cihak

D. F.

(2005). Using the Expressive Writing program to improve the writing skills of high school students with learning disabilities. Learning Disabilities Research & Practice, 20, 175–183. doi:10.1111/j.1540-5826.2005.0131.x

39.

Walker

B. D.

Shippen

M. E.

Houchins

D. E.

Cihak

D. F.

(2007). Improving the writing skills of high school students with learning disabilities using the Expressive Writing program. International Journal of Special Education, 22, 66–76.

40.

Wolery

(2013). A commentary: Single-case design technical document of the What Works Clearinghouse. Remedial and Special Education, 43, 39–43.