Parker, D. C., Burns, M. K., McMaster, K. L., Al Otaiba, S., & Medhanie, A. (2017). Examining the relations between instructional-level data and intervention response in early writing. Assessment for Effective Intervention. Advance online publication. doi:1534508417731999
National assessments of student performance indicate that three-fourths of students are not proficient in writing and that two-thirds of students are not proficient in reading (National Center for Education Statistics, 2012; 2013). These findings indicate a need to address reading and writing difficulties preventatively. However, as in reading, not all evidence-based writing interventions are likely to benefit all students because students vary in their level of initial writing performance.
Parker, Burns, McMaster, Al Otaiba, and Medhanie (2017) analyzed data from a larger literacy intervention in which students received a yearlong reading intervention that included a writing intervention (Al Otaiba et al., 2014). The focus of the present study was on writing intervention. The authors analyzed longitudinal data from an 8-week writing intervention for 147 first-grade students assigned to either a Tier 2 or Tier 3 intervention. They first determined growth patterns during the 8-week intervention period. They then examined the data for associations between growth patterns and students’ initial level of writing ability based on the students’ instructional level (described below): (a) independent, (b) instructional, and (c) frustration.
A number of theoretical models of reading comprehension are described in the literature. One of the most robust ones is the simple view of reading, which states that students must balance code- and meaning-focused approaches to comprehend text. These approaches refer to decoding words accurately and fluently and integrating information in the text with background knowledge. A similar theoretical model of writing is the not-so-simple view of writing. This model holds that writing requires a balance among lower-level transcription skills and higher-level cognitive skills. Skills such as handwriting and spelling provide a foundation for transcription skills. Higher-level skills involved in composition require text generation, planning, text organization, and self-regulation among other writing processes. Evidence-based writing interventions have identified procedures for improving transcription and composition. However, because students might struggle with one or more skills, not all standard protocol interventions are likely to benefit all students because students have different needs.
To match interventions to student needs, Parker et al. (2017) created a task-skill difficulty level referred to as instructional level. Learning is optimized when the instruction falls within a student’s zone of proximal development (i.e., the task is neither too easy nor too difficult, but challenges the student). This match between student ability and task demand is the instructional level. When students demonstrate proficient skills before instruction, there is little room to grow (i.e., the task is too easy and students are at the independent level). Finally, when students' skills start too low and the instruction does not adequately address their learning needs, the students are at the frustration level (Parker, McMaster, & Burns, 2011). Thus, matching interventions to student needs may help determine when writing interventions should focus on higher-level skills such as planning and organizing or lower-level skills such as handwriting and spelling. In other academic domains, instructional levels are determined by sampling behavior for a short period of time (e.g., reading rate or digits correct per minute), and the resulting data are sufficiently reliable for instructional decisions in reading and math (Parker et al., 2011). Less is known about the adequacy of instructional decisions in writing.
In this study, the number of correct word sequences (CWS) was used to measure writing performance in a 3-minute testing period. CWS is a curriculum-based measure of transcription and consists of any two adjacent, correctly spelled words that are also used correctly within the context of written English. Students completed the writing probe across seven time points during the intervention period. Importantly, in the present study, the writing intervention was conceptualized as aligning with the independent level (i.e., students who wrote more than 14 CWS at the beginning of the intervention) because the intervention focused on idea generation, a higher-order writing process.
What growth patterns do students follow during a brief standard protocol writing intervention?
Overall, students with higher initial writing ability grew faster than those with lower initial ability. The following three growth profiles were found using growth mixture modeling:
To what degree are observed individual growth patterns related to instructional-level data?
The authors determined the agreement between instructional level and growth patterns using Cohen’s kappa, or the proportion of agreement after correcting for chance. To this end, students were at the instructional level if they wrote 8 to 14 CWS at the beginning of the intervention, at the independent level if they wrote more than 14 CWS, and at the frustration level if they wrote fewer than 8 CWS. The authors found a low-to-moderate significant agreement between methods (kappa = 0.38). Key findings are as follows:
The overall pattern of results shows that students with better writing skills at the beginning of the intervention made the largest gains in writing. This finding is consistent with a “fan-spread” growth pattern commonly found in educational literature. This growth pattern is also referred to as the Matthew effect and results in a widening achievement gap between students with low vs. high transcription skills over time. The present study shows some promise for informing the degree to which student skills align with the intervention focus. However, it does not directly address how initial status data might prospectively inform the balance between lower- and higher-level skills because to do so would require an experimental design with multiple treatments (e.g., writing protocols that match students' instructional level). Furthermore, the findings of this study also need to be interpreted in the context of the following limitations: (a) Reading intervention occurred concurrently with writing interventions, (b) students were assigned to an intervention based on reading assessments rather than writing assessments, and (c) results did not distinguish between Tier 2 and Tier 3 students.
Despite these limitations, this study has important practical implications. Educators often struggle to balance lower- and higher-level skills in writing instruction. The results of this study suggest that educators working with struggling writers may benefit from using assessment data to understand skill deficits and consider ways to differentiate targeted interventions. Students with the highest transcription skills appeared to benefit from a writing intervention focused on text generation, whereas students with the lowest transcription skills seemed to need an intervention focused more on accurate and fluent transcription skills such as handwriting and spelling.
Al Otaiba, S., Connor, C. M., Folsom, J. S., Wanzek, J., Greulich, L., Schatschneider, C., & Wagner, R. K. (2014). To wait in Tier 1 or intervene immediately: A randomized experiment examining first grade response to intervention in reading. Exceptional Children, 81, 11–27.
National Center for Education Statistics. (2012). The nation’s report card: Writing 2011 (NCES 2012–470). Washington, DC: Institute of Education Sciences.
National Center for Education Statistics. (2013). The nation’s report card: A first look: 2013 mathematics and reading (NCES 2014–451). Washington, DC: Institute of Education Sciences.
Parker, D. C., McMaster, K. L., & Burns, M. K. (2011). Determining an instructional level for early writing skills. School Psychology Review, 40(1), 158–167.