Setting
The study took place in nine schools across three school districts in the southwestern United States. Six schools were in urban areas, and the authors note that of those six, three were in a large urban school district, and three were in a small, affluent urban school district. The other three schools were in a “near-urban” school district. The mean enrollment of the participating schools was 585 students (SD=119.8, range=440-728). The proportion of students who passed the reading section of the state accountability exam in fourth grade ranged from 34% to 99% (M=75.2%, SD=22.5%).
Study sample
Study participants included fourth- and fifth-grade students with severe reading difficulties. Of the analytic sample, about 52% of students were male and 48% were female. About 21% of students had limited English proficiency. The analytic sample was mostly White (51%), followed by Black (42%), American Indian (3%), Asian (2%), and other or multiple races (2%). About 14% of students in the analytic sample were in special education. Within the participating schools, the proportion of students qualifying for free or reduced-price lunch ranged from 2.6% to 96.1% (M=44.4%, SD=37.5%).
Intervention Group
The study examined the effectiveness of a reading intervention for students struggling with reading. The intervention was delivered to small groups of three to six students in 30- to 45-minute lessons. The intervention was delivered five days per week from October to April. On average, students participated in 68 lessons. The intervention proceeded in three phases. During the first phase, lessons focused on word study and reading fluency. Twenty minutes each day were devoted to word study including systematic decoding instruction and practice with word patterns and sight words. The words and text that were covered in the lesson progressed over time to more difficult words and text. Students started by learning words and then went on to work with the words in the context of phrases, sentences, and longer text. Students also practiced spelling and took spelling quizzes. Ten minutes per day in phase 1 was devoted to fluency. Students worked with QuickReads texts during this time. This part of the lesson included the introduction of keywords related to the text's main idea, repeated reading of the text in different formats, and using the keywords in a summary of the passage. Phases 2 and 3 focused on reading expository, narrative, and hybrid texts. For three out of the five lessons, students worked with stretch texts, which were one grade level above their current reading level. These texts were similar in terms of vocabulary and background knowledge to the grade-level texts that students had previously read. Students set goals, read the text, answered comprehension questions, and then evaluated their goals. During the other two lessons, students worked on fluency and word study. Fluency instruction was similar to phase 1. Word study included systematic instruction in morphology for 10 to 15 minutes and automaticity instruction that focused on words and sentences covered in the first phase. AiMSweb was used for progress monitoring throughout the intervention.
Comparison Group
Students in the comparison group participated in business-as-usual instruction. Approximately two-thirds of comparison students received at least one additional reading intervention that was provided by the school. The nature of these interventions varied and included use of commercially available curricula, computer-based instruction, and teacher-designed lessons, which focused on word reading, fluency, comprehension, and preparation for high-stakes assessments.
Support for implementation
The tutors participated in 18 hours of professional development (PD). Of those PD hours, ten hours occurred prior to the implementation of the intervention, and eight hours occurred after lesson 40. Also, every two to three weeks, tutors participated in staff development meetings and on-site feedback and coaching. Tutors audio-recorded all instructional lessons, and a minimum of 20 percent of lessons per tutor, per block (i.e., lessons 1-40, lessons 41-80, lessons 81-110) were randomly selected for fidelity coding. The fidelity coding was conducted by four trained research assistants. Coders assigned a score using a 4-point Likert-type rating scale to evaluate implementation fidelity for specific intervention components as well as for global quality (i.e., adherence to principles of teacher instruction and classroom management) and global fidelity (i.e., a holistic evaluation of implementation and success of the lesson).