Grunér, S., Östberg, P., & Hedenius, M. (2018). The compensatory effect of text-to-speech technology on reading comprehension and reading rate in Swedish schoolchildren with reading disability: The moderating effect of inattention and hyperactivity symptoms differs by grade groups . Journal of Special Education Technology , 33 (2), 98–110.

Journal Article

Grunér, S., Östberg, P., & Hedenius, M. (2018). The compensatory effect of text-to-speech technology on reading comprehension and reading rate in Swedish schoolchildren with reading disability: The moderating effect of inattention and hyperactivity symptoms differs by grade groups. Journal of Special Education Technology, 33(2), 98–110.


Attention problem; Elementary; International (non-U.S.); Learning disabilities; Middle school; Multiple ages; Oral delivery; Reading; Student reads aloud (to self); Text-to-speech device/software




A computer-administered assessment with both text passages and test items was presented in two versions for comparison: oral delivery with text-to-speech by computer-simulated voice along with screen display marking as the test was read, versus student read-aloud (to self) from the screen.


Students with reading disabilities—31 students in grades 3–5 and 18 students in grades 6–9—participated. The participants, who attended school in Sweden, had been referred by school nurses or family doctors to a clinic addressing speech-language needs. Two participants had also been diagnosed with attention-deficit hyperactivity disorder (ADHD), and three had been referred for consideration for possible ADHD diagnoses. Demographic information such as sex (male/female) and age were also reported, and grade levels were identified for each participant. Other data pertinent to the clinic referral served as independent information essentially for screening purposes regarding the participants. The Strengths and Difficulties Questionnaire (SDQ; Goodman, 1999; Smedje et al.,1999)—consisting of scales for hyperactivity/inattention, emotional problems, behavioral/conduct, peer problems, and prosocial behavior—was completed for all participants. In addition to the reading decoding and comprehension tests that served as dependent variables, documentation of broader language skills were available: Swedish versions of the Boston Naming Test (BNT; Kaplan et al., 1983) and the Test for Reception of Grammar (TROG; Bishop, 2003). Reading decoding skills were also documented, with a standard set of measures (Elwér et al., 2009; Jacobson, 2014; Johansson, 2004) used in Sweden.

Dependent Variable

A common computer-administered comprehension assessment, Logos (Høien, 2007), including reading and listening comprehension, was employed for each of the two grade ranges. The test used grade-level informational text paragraphs interspersed with reading comprehension items; the grades 3–5 test had five fictional content paragraphs with three items each, totaling 15 items, and the grades 6–9 test had five factual content paragraphs with four test items each, for a total of 20 items. Test items called for short-answer responses, which test-takers made verbally into a microphone for documenting their responses. Test-takers' times for completing the readings were tracked, providing reading rate data. After participants completed assessments in both versions (oral delivery by TTS and student read aloud to self), students were asked in brief interviews to respond to two routine questions on their preference between versions and their previous TTS experience, which were also apparently reported orally and documented in participant recordings. Performance score means were reported for the participant group as a whole as well as for each grade band, grades 3–5 and grades 6–9, in both testing conditions.


All participants increased their reading rate with text-to-speech (TTS) and 71% of students increased their reading comprehension with TTS. When examined by grade bands, TTS had a significant average benefit—that is a "compensatory effect"—on reading rates for participants in each grade band. TTS showed benefits in comprehension performance for the participants in grades 3–5; however, there was no significant mean difference in comprehension for participants in grades 6–9 in the TTS condition versus the student-reads-aloud (to self) condition. An additional analysis indicated that TTS benefited participants with SDQ scores that were high for ADHD symptoms. Grade-band analyses yielded a more complex picture. For participants in grades 3–5 with low ADHD indicators (on SDQ), TTS supported better comprehension, but TTS use did not benefit these younger students with high ADHD indicators; in fact, TTS use produced lower comprehension scores on average for them. By contrast, TTS use had a compensatory effect for participants in grades 6–9 with high ADHD indicators. TTS use did not have any significant effects on reading rates related to SDQ scores on ADHD symptoms. For 14 participants—6 of 31 younger students and 8 of 18 older students—TTS use resulted in lower comprehension scores than when the participants read aloud to themselves. These participants had higher comprehension performance overall than the participants who benefited from TTS use; a closer view indicated that this pattern applied only to the younger participants. For the older participants who dropped in comprehension performance with TTS, their ADHD indicators were lower than those who benefited from TTS use. About 75% of all participants preferred the TTS condition, and 25% preferred to read the test aloud to themselves; grade-band analyses yielded equivalent preference proportions. About 80% had little or no previous experience using TTS, and 20% reported previous regular use of TTS. About 70% of those regularly using TTS preferred it over the other condition; about 82% of those with little or no TTS experience preferred the TTS condition. For the most part, participants who preferred TTS also had larger compensatory effects on comprehension; however, about 50% of participants with lower comprehension scores when using TTS had reported that using TTS was easier than reading aloud to themselves.