Writing characteristics of students with learning disabilities and typically achieving peers: A meta-analysis

There is a general consensus that writing is a challenging task for students with learning disabilities (LD). To identify more precisely the extent and depth of the challenges that these students experience with writing, the authors conducted a meta-analysis comparing the writing performance of students with LD to their typically achieving peers. From 53 studies that yielded 138 effect sizes, the authors calculated average weighted effect sizes, showing that students with LD obtained lower scores than their peers on the following writing outcomes: writing quality (–1.06); organization (–1.04); vocabulary (–0.89); sentence fluency (–0.81); conventions of spelling, grammar, and handwriting (–1.14); genre elements (–0.82); output (–0.87); and motivation (–0.42). Implications for research and practice are provided based on these findings. Good writing is essential to success. At school, it is used to assess what students know, support students’ comprehension of text, and facilitate learning of content material (Graham & Hebert, 2011). At work, writing is a common activity in white- and blue-collar jobs, and writing proficiency is increasingly used to make decisions about who gets hired and promoted (National Commission on Writing, 2004). At home, writing is pervasive, through constant e-mailing, texting, tweeting, and posting as well as correspondence with friends, family, and others. Because of the importance of writing, students who experience difficulty mastering it are placed at a disadvantage. Students with disabilities, including students with learning disabilities (LD), are at particular risk. There is a general consensus that students with LD evidence greater challenges with writing than their typically achieving counterparts (Graham, Harris, & McKeown, 2013), and data from the most recent National Assessment of Educational Progress (U.S. Department of Education, 2011) confirm this belief. Across all students identified with a disability, only 5% achieved writing proficiency, and 60% of students with disabilities failed to meet basic writing achievement levels. A large percentage of these students were youngsters with LD. Why is writing so challenging for students with LD? First, writing is a very complex process, making it a challenging task for all students (Graham & Harris, 2000). It requires the orchestration of handwriting, typing, spelling, and sentence construction skills that allow composing to take place; strategies for planning, evaluating, monitoring, drafting, and revising text; topic, genre, linguistic, and semantic knowledge for creating meaning; and the motivational aspirations to put these skills, strategies, and knowledge into play. Second, students with LD experience difficulty with general cognitive and affective processes that underlie skilled writing. When compared with their typical peers (Swanson, Harris, & Graham, 2013), students with LD have weakened memory skills as well as difficulties with executive functioning and cognitive monitoring skills. These challenges may affect their access to needed writing knowledge and strategies. Students with LD also experience difficulty mastering basic skills such as phonological awareness in reading (Shankweiler, Lundquist, Dreyer, & Dickinson, 1995). The mechanisms responsible for such difficulties may in turn hinder the development of basic writing skills. Students with LD further exhibit a lack of academic self-confidence (Bryan, 1991) and may be less motivated to exert sustained efforts with challenging tasks like writing. The application of meta-analysis to the available literature allowed us to determine the direction and magnitude of difference in the writing of these two groups of students, providing a more systematic picture of the breadth and depth of writing challenges faced by students with LD. Previous reviews of research examining the writing performance of students with LD have confirmed that writing is challenging for these youngsters (e.g., Graham et al., 2013; Graham, Harris, MacArthur, & Schwartz, 1991; Lynch & Jones, 1989; Newcomer & Barenbaum, 1991), but these reviews have not provided a comprehensive summary of the research. Graham et al. (1991, 2013) conducted selective reviews of the literature for students with LD in Grades 1 to 12, focusing primarily on their own research. Lynch and Jones (1989) concentrated on elementary students with LD but limited their systematic review to 5 years. Similarly, Newcomer and Barenbaum (1991) examined studies of the writing of children and adolescents with LD but limited their review to the 1980s. There appears to be no systematic, comprehensive review of the literature examining the writing of students with LD in Grades 1 to 12. Our study addresses this gap in the literature as it involved a comprehensive meta-analysis of studies comparing the writing of LD students and their typical peers in Grades 1 to 12. The application of meta-analysis to the available literature allowed us to determine the direction and magnitude of difference in the writing of these two groups of students, providing a more systematic picture of the breadth and depth of writing challenges faced by students with LD. We were interested in students’ text production skills, self-regulation strategies for carrying out writing processes, knowledge about writing, and motivation to write (Deane, 2013). If students with LD experience difficulties with text production skills (e.g., sentence fluency, handwriting, spelling, and grammar) when compared to typical peers, this should affect how well and how much they write. Until text production skills can be executed efficiently, they exert a toll on writing. For instance, students may forget plans and ideas they are trying to hold in working memory as they stop to think about how to spell a word. If such difficulties are common, they may negatively affect the length, organization, and quality of students’ text (Graham & Harris, 2000). Difficulties with text production skills may further affect use of strategies and knowledge. If text production skills require considerable cognitive resources, writers may minimize use of strategies for planning, evaluation, and revising as these skills are also resource intensive. They may also minimize their use of available writing knowledge. For example, they may only choose words they know how to spell correctly. Finally, difficulty with text production may result in students’ viewing themselves as poor writers, reducing motivation for writing. Further, difficulties acquiring self-regulation writing strategies, knowledge about writing, or positive dispositions about writing should negatively affect text length, organization, and quality. For example, self-regulation strategies for planning and revising text provide mechanisms for gathering, organizing, and evaluating possible content for writing as well as evaluating and revising the text produced (Graham & Harris, 2000). Knowledge about how to write, including genre knowledge, provides schemas for creating and structuring text (Deane, 2013). Motivational dispositions affect what writers do as writers with an “I can do” attitude are more likely to plan, set challenging goals, and persist as writers. We predicted that students with LD would evidence writing challenges in each of the areas described previously when compared to their typical peers. They experience difficulties acquiring basic skills, problems with executive functioning and cognitive monitoring, and reduced motivation (Swanson et al., 2013) that are likely to undermine their performance to write. One advantage of conducting a meta-analysis is that relationships between and among specific study features and study outcomes can be examined when there are enough studies and the variance between effect sizes (ESs) is greater than sampling error alone. Thus, when both of these criteria were met, we conducted moderator analyses to determine if the overall effect for a writing outcome varied by study quality, grade level (elementary vs. secondary), year of publication, and genre. We predicted that study quality would be related to magnitude of the overall effect (as in Moher et al., 1998). Studies of higher quality should produce smaller effects as methodological rigor should minimize inflated outcomes or erroneous findings (Pressley, Graham, & Harris, 2006). Due to the Matthew Effect (Stanovich, 1986), where stronger students make greater academic gains over time than students who experience challenges academically, we anticipated that the writing challenges faced by students with LD would be more pronounced at the secondary level than at elementary grade levels. Because studies in this review spanned 40 years, it is possible that improvement in measures and study techniques as well as changes in the LD population may have systematically affected study outcomes. No prediction was made for this analysis, but we did expect that writing genre would be related systematically to study outcomes as even struggling writers’ performance can vary from one type of text to the next (Graham, Hebert, Sandbank, & Harris, 2016). In summary, this meta-analysis asked: Are there differences in the writing of students with LD and their typically achieving peers? To answer this question, we examined differences in these students’ written products, text production skills, self-regulation strategies, writing knowledge, and writing motivation.