Meyer, N. K., & Bouck , E. C. (2017). Read-aloud accommodations, expository text, and adolescents with learning disabilities . Learning Disabilities: A Multidisciplinary Journal , 22 (1), 34–47. https://doi.org/10.18666/ldmj-2017-v22-i1-7932

Performance was compared when using oral delivery presented live by a test administrator versus through text-to-speech computer software (Natural Reader from NaturalSoft Limited, 2012). The researchers noted that both testing conditions provided individual administration and quiet alternate setting (e.g., in the resource room and not in the general education classroom).

Four grade 7 students with learning disabilities participated in this alternating treatment design study. Additional demographic data, such as gender, and other academic information including circumstances of identification with disabilities, pre-screening reading comprehension, and IEP accommodations and academic services, were reported. All participants had decoding and comprehension skills that were at least two grade levels below their current placements. The researchers also provided contextual information about the junior/senior high school in a Midwestern city (U.S.), including student population demographics such as race/ethnicity.

Participants responded to a series of common comprehension questions, based on a rubric designed by the researchers, inquiring about main idea and supporting details in each reading passage, and a set of passage-specific questions. Passages were connected to grade 7 content, and drawn from a commercially-available reading curriculum (Pauk, 2000); the passages were expository, containing fact-based content in science and social studies. Students each participated in about 6 baseline sessions, 5 intervention sessions in each accommodation condition, additional "best treatment" (as determined by intervention performance) sessions, and a few subsequent generalization/maintenance sessions. The generalization passages were new academic content from students' classroom materials on common topics in science and social studies; the reading tasks were focused on reading comprehension and not knowledge in academic content. Participants answered the comprehension questions orally to the researchers, and their responses were scribed during the testing sessions. Participants' performance data, which were the percentage of correct answers from each set of passage questions, were analyzed based on percentage of nonoverlapping data (PND) and improvement rate difference (IRD) metrics. Participants were screened for their reading performance using the Woodcock Reading Mastery Tests-Revised (1987), including Word Identification, Word Attack, and Passage Comprehension. All students also participated in two social validity interviews—one after training and another after maintenance sessions, asking about their opinions of using the two different oral delivery versions of the comprehension questions, including their preferences between them if any. The researchers also inquired about the participants' comprehension challenges by asking brief interview questions of the students' teacher. 

Performance data were reported graphically for each participant, and individual descriptions of student performance were provided. Overall, the researchers concluded that oral delivery did not support any significant performance improvements for all four participants; they described the effect sizes of both forms of oral delivery for three participants as "small to questionable" (p. 44) and stated "only one student experienced modest effects for both types of read-alouds (i.e., Cole)" (p. 44). Further, according to the improvement rate difference (IRD) data, the participants did not score significantly differently when using one oral delivery format over the other, and as the researchers noted, "the data in both conditions were variable, with a great deal of overlap" (p. 44). They also noted that the percentage of nonoverlapping data (PND) analyses indicated that oral delivery in-person by a test administrator yielded better scores for three of the students, and text-to-speech supported one student (Kara) better; however, these differences were judged to be no greater than modest. Early in the study, participants all expressed their expectation that text-to-speech would likely help them more; however, by the end of the study, three of the four participants preferred in-person oral delivery over text-to-speech. The researchers stated that the participants' special education teacher "was skeptical of students using TTS [text-to-speech] appropriately, although she noted that it might provide more independence for the students as well as improve their efficiency" (p. 44). The researchers offered the observation that oral delivery might not be beneficial as a sole intervention for students with reading-based disabilities, and recommended additional instructional interventions need to be provided. Limitations of the study were reported, and future research directions were suggested.