The Effects of Text-to-Speech on Reading Outcomes for Secondary Students With Learning Disabilities

Abstract

This study investigated the effectiveness of text-to-speech (TTS) on the outcomes of reading comprehension and oral reading fluency (ORF) for four secondary students with learning disabilities. The researchers used a single-case A–B–A–B withdrawal design to evaluate the effectiveness of TTS on reading outcomes. All participants scored higher on reading comprehension after using TTS when reading instructional passages and maintained the skills for 4 weeks. Results on participants’ ORF also indicated an increased level of words read per min at the end of each accommodation condition. Comparison of pre- and posttest achievement on the Lexile assessment showed that two of the four participants increased their reading scores. Major findings are discussed with implications for practice and recommendations for future research to increase the use of TTS in the classroom.

Keywords

assistive technology learning disabilities legislation secondary text-to-speech

Reading is a necessary and critical skill. The acquisition of reading substantially impacts educational outcomes, employment success, and personal and professional growth (Strangman & Dalton, 2005). In the United States, roughly 6 million secondary students are reading at a level far below their grade. The National Assessment of Educational Progress (NAEP) from 2017 indicated that 74% of eighth graders do not have proficient-level reading skills even though the data showed an overall slight increase of reading achievement since 2015. However, it should be noted that the growth rate is with students who were higher achieving, whereas students with disabilities remained statistically the same. The NAEP reading achievement scores for students with disabilities in 12th grade across the nation have not increased since 2009, the earliest available data for students with disabilities at that grade level. In fact, the gap between 12th graders with and without disabilities has increased by 3 points since 2009, indicating that those with disabilities are continuing to fall behind their peers without disabilities (U.S. Department of Education, 2015, 2009, 2017).

Students With Learning Disabilities (LDs)

Several legislative actions have focused on evidence-based practices and inclusion of students with disabilities in the general education curriculum; yet as indicated by NAEP results, reading scores have not improved for many students with disabilities over the years. These results are especially concerning for students with LDs who are the highest population of students receiving special education services at 35% (Snyder, de Brey, & Dillow, 2016; Wanzek, Otaiba, & Petscher, 2014). Approximately 80% of students with LD exhibit deficits in the area of reading (Cortiella & Horowitz, 2014). In addition, 90% of students with LD are not able to read material independently (Vaughn & Wanzek, 2014). Students who have difficulties making meaning from text are likely to encounter postschool challenges, such as being unemployed, earning lower incomes, and exhibiting poor health as adults (Snyder & Dillow, 2013). In order to reduce the negative postschool outcomes associated with reading difficulties, addressing adolescents’ lack of reading progress in our high schools is imperative.

Students with LD often experience several years of reading difficulties that involve deficits in making meaning from text (Cortiella & Horowitz, 2014). Achieving success in school requires proficient reading skills to understand all content areas (Vaughn & Wanzek, 2014). This gap widens in high school as students struggle to gain information from text that is necessary for learning (Saenz & Fuchs, 2002). To prevent the achievement gap from further widening between special and general education students, additional types of reading accommodations need to be implemented. Technological accommodations have been known to increase academic outcomes for secondary students with LD (Stetter & Hughes, 2011).

The National Educational Technology Standards and the Council for Exceptional Children Technology Specialist established standards to train teachers on the use of assistive technology (AT) when instructing students with disabilities (Parette & Peterson-Karlan, 2010; Smith & Allsopp, 2005). Even with these guidelines, many teachers are not routinely using technology to make content accessible (Okolo & Diederich, 2014). Adding to this dilemma is that classroom curricula are primarily based in print (Rose, Meyer, & Hitchcock, 2011). Given this, students who are able to read the text may benefit more than students who struggle to read. This may lead to a Matthew Effect that causes the word-rich to get richer and the word-poor to get poorer (Stanovich, 1986). To aid in addressing this issue, teachers need to boost students with disabilities’ motivation to read. Technology is highly motivating to secondary students with LD and can potentially increase the amount of content they learn and read (Anderson-Inman, 2009; Bouck, Flanagan, Miller, & Bassette, 2012).

AT

The availability of AT for individuals with disabilities has increased dramatically since the 1970s (Poel, 2007). In a recent state survey, 67% of teachers reported that they believed AT increased student access to curriculum and outcomes, but only 33% of respondents could make print accessible on the computer for their students (Okolo & Diedrich, 2014). When asked what contributes to using technology, more than half of the teachers identified the severity of the students disabilities, as in, the more severe a students’ disability, the more likely they were to use technology. The teachers in this study also identified staff knowledge as a barrier (70%) along with availability of technology (61%) and funding (61%). Despite increased AT device presence, teachers are not typically using AT in the classroom, especially classrooms with students who have a less severe disability (e.g., LD). Unfortunately, many teachers perceive AT for a certain type of disability, have a lack knowledge of how to use AT, and identify obtaining devices as problematic.

The Tech Act, reauthorized in 2004, intensified the previous legislation to promote the access, knowledge, and obtainability of AT among students with all disabilities to access curriculum. Prior to that in 1998, The Tech Act only defined AT as any piece of equipment or product system, whether acquired commercially, off the shelf, modified, or customized, that is used to increase, maintain, or improve the functional capabilities of individuals with disabilities (Assistive Technology Act of. 1998). Additionally, Individuals With Disabilities Act (2004) requires every individualized education program (IEP) team to consider AT for students with disabilities during annual meetings (Parette & Peterson-Karlan, 2010). If the IEP team determines AT is needed, an AT evaluation is conducted and listed as a related service or supplementary aid in the student’s IEP (Parette & Peterson-Karlan, 2010). The intention of this mandate along with the discussion of Universal Design for Learning was to increase the consideration of using AT more frequently to help students access curriculum in the classroom, regardless of the severity of the disability. However, these legislative actions did not increase AT in the classrooms as indicated by Okolo and Diedrich’s study in 2014.

Benefits of Text-to-Speech (TTS)

To assist secondary students in accessing curriculum, students with reading disabilities oftentimes need compensatory tools (e.g., TTS). TTS is an AT that can compensate for reading difficulties and can increase access to text (Stodden, Roberts, Takahashi, Park, & Stodden, 2012). AT used as a compensatory tool is defined as any tool permitting a person to complete a task where without the tool, the person is unable to perform the task at the expected level (Courtad & Bouck, 2013). Generally, technology has shown to promote independence and self-worth and to increase motivation and productivity among students with LD (Edyburn, Higgins, & Boone, 2005). More specifically, technological speech–synthesized accommodations have resulted in reading gains for students with LD (Moorman, Boon, Keller-Bell, Stagliano, & Jeffs, 2010).

Following the use of TTS, Gruner, Ostberg, and Hedenius (2017) found that students improved reading rates, increased time spent on reading, and performed better on comprehension measures. They also noted that both elementary-aged and secondary students improved their reading rates when using TTS. However, only elementary school students showed improvement in reading comprehension. Furthermore, Moorman, Boon, Keller-Bell, Stagliano, and Jeffs (2010) found that the lowest readers benefited the most from TTS in comprehension. Specific TTS features have been known to enhance students’ engagement and outcomes (Wood, Moxley, Tighe, & Wagner, 2018). When individualizing TTS through the use of embedded features (e.g., reading rate, voice type, and highlighting), students with LD exhibited educational gains including increased reading rates and comprehension (Moorman et al., 2010; Tian & Okolo, 2007; Young, 2013). Research also showed that when computerized reading speed was set at a faster rate than present oral reading fluency (ORF), students increased their comprehension (Coleman, Carter, & Kildare, 2011).

Reading difficulties for secondary students who struggle to make meaning from text will likely continue beyond the school years in their postschool life. TTS is an accommodation that is nearly ubiquitous, across operating systems, platforms, and applications (Stodden et al., 2012); yet classroom teachers reported not using AT on a regular basis (Okolo & Dietrich, 2014). As Okolo and Dietrich wrote, “…research about the impact of AT remains limited. Few papers in an educational journal end without a plea for more research” (p. 18).

Purpose

In order to answer the professional call for more research on the impact of AT for students with LD and add to previous research indicating positive outcomes for students with disabilities, this study uses AT as an accommodation for secondary students with LD who struggle to make meaning from text in a traditional print form. Specifically, this study evaluated the effectiveness of TTS on the reading comprehension of high school 9th graders with LD in a self-contained class. The first author (i.e., a special education teacher) used the reading accommodation of TTS, Kurzweil 3000, to assist with reading material for students with LD. She monitored participants’ performance in reading comprehension through curriculum-based assessments (CBAs) to ensure improvement before measuring the effects on outcomes in maintenance sessions. The results provided additional empirical evidence on the effects of TTS when used as an accommodation for secondary students with LD.

The research questions included the following: (a) Does the reading comprehension of freshman students with LD increase when using TTS as measured by CBAs? (b) Will students maintain the use of TTS following intervention? (c) Does ORF increase after students use TTS? (d) and To what extent do the students generalize reading comprehension to pre and post universal screening assessments (e.g., Lexile)?

Method

Participants

The first author selected participants based on purposive and convenience sampling. She recruited potential participants for this study if they met the following criteria: (a) were identified as having a LD in reading, (b) received services in a 9th-grade self-contained English class, and (c) had a 95% attendance rate at the start of the study. In this district and state, students qualified as having LD after progressing through a response to intervention (RtI) model, demonstrating a lack of progress at level three, and having a discrepancy between their cognitive abilities and their performance. The self-contained English class was required for all students with LD in reading who had a Lexile score in the bottom 25% locally or below 900 and had an IEP with a reading comprehension goal. This class had a total of enrollment of 11 students. The students in the self-contained English class were also enrolled in a self-contained reading class using a different curriculum package, READ 180 by Houghton Mifflin Harcourt (n.d.). Four of the 11 students meeting the inclusion criteria provided informed assent and parent permission (see Table 1 for participant demographics). Only three students completed the entire study through maintenance and generalization. Vincent left the study early because of a change in his class schedule.

Table 1.

Participant Demographics.

Name	Age	Initial Eligibility	Ethnicity	Used TTS Previously	Initial Lexile Scores	Percentage in General Education
Gender	Age	Primary and Secondary Characteristic	Ethnicity	Used TTS Previously	Initial Lexile Scores	Percentage in General Education
Dianna	15	11/2012	African American	No	772	40–79
(F)		SLD/ADHD
Donald	14	4/2008	African American	No	872	40–79%
(M)		SLD
Jack	14	12/2009	African American	No	186	Less than 40%
(M)		SLD/AU
Vincent	15	5/2011	African American	No	864	40–79%
(M)		SLD

Note. ADHD = attention deficit/hyperactivity disorder; AU = autism; SLD = specific learning disability.

Setting

The study took place in a large, Midwest public high school with an estimated overall student enrollment of 3,500 and approximately 15% of the students receiving special education services. Forty percent of these students receiving special education services had LD. Twenty percent of the overall student population received free and/or reduced-price lunch. Research activities occurred in a computer lab close to the students’ self-contained English classroom. The computer lab had at least 15 desktop computers with the TTS program, Kurzweil 3000 (Kurzweil Education, 1996), installed on each computer. All students in the class (n = 11) used TTS during the scheduled self-contained English class to access the curriculum; however, the first author only collected data on the four students participating in this study. Students sat at least one seat away from their peers to minimize distractions. Each TTS session lasted one class period (i.e., 48 min) for 3 times per week during the scheduled English class time in the early morning.

Materials

Classroom textbook

The district adopted a complete curriculum called Hampton-Brown Edge: Reading, Writing, & Language (Moore, Short, Smith, & Tatum, 2014). This curriculum contains four reading levels for adolescent students. All participants in this study used Level A based on students’ reading levels in the bottom 25th percentile on Lexile scores locally. Before the study began, the first author selected 29 fiction stories from the The Edge Level A (Moore et al., 2014). In addition, she used The Edge curricular comprehension assessments for each story. The comprehension assessments consisted of 15 multiple-choice questions, including eight vocabulary, four literary analysis, and three reading comprehension questions.

TTS

Prior to beginning the study, participants attended a training session conducted by the first author on how to use Kurzweil 3000. During this training sessions, students were allowed to personalize the settings for highlighting text and choose the voice reading the text. The first author choose the setting for the read aloud rate (i.e., how fast the TTS read). These individualized settings became the default each time a participant logged into Kurzweil 3000 to access the stories.

Dependent Measures

Reading comprehension measures

For this study, the first author used a reading comprehension measure to evaluate the effectiveness of TTS among students with reading deficits within a withdrawal design study as recommended by Perelmutter, McGregor, and Gordon (2017). After students read each fiction story, the reading comprehension measures included the CBAs from the Edge assessments (Moore et al., 2014). The first author scored the CBAs by grading the percentage of correctly answered questions out of the 15 total questions for each story. Additionally, she used the Scholastic Reading Inventory (SRI) assessment to obtain pre- and posttest Lexile score (Lexile, n.d.). The SRI is a computerized reading comprehension assessment that lasts approximately 30 min. It is used to obtain a reader’s Lexile score ranging from 0 to 1500 and as a universal screener to determine placement in English courses. The SRI recommends placing students into categories to identify their reading needs.

ORF

The first author also used an ORF measure to determine whether students increased their words read per minute after auditory and visual exposure to text. The RtI specialist employed at the high school selected the ORF passages from Aimsweb (Pearson, 2012). Each student read three paper-based 1-min 8th-grade passages while the teacher documented the number of words read correctly. Each ORF passage contained a different topic, so students did not repeatedly read the same passage. After participants read three passages in one session, the teacher averaged the scores to determine ORF scores for that date and time. The teacher conducted ORF assessments in this manner 6 different times throughout the study: prebaseline, after each condition, and after the maintenance condition. She administered a total of 18 ORF assessments during the first half of the school year.

Social validity

The classroom teaching assistant administered a teacher-created 8-item Likert-type scale survey to each participant in the instructional classroom to gather social validity information on the procedures and outcomes of the TTS accommodation (survey available upon request). The survey asked participants to rate their levels of satisfaction on a 5-point scale (i.e., 1 = strongly disagree; 5 = strongly agree) regarding the use of TTS. Because the teacher created the survey at the participants’ readability levels, the participants read the survey on their own. Mean scores on each item were reported.

Experimental Design

A single-case A–B–A–B withdrawal design (Gast & Ledford, 2014) was used to evaluate the effectiveness of the TTS accommodation on the students’ reading comprehension after reading passages from the classroom text. Each student served as her or his own control as this design allowed for intra- and intersubject replication. Students changed conditions after they completed a predetermined number of CBAs (i.e., five, seven, six, and seven in the first baseline, first accommodation, the second baseline, and the second accommodation conditions, respectively). The number of sessions in each condition varied as a way to control for cyclical threats to internal validity in a withdrawal design.

Procedures

Training

Before the baseline condition began, students engaged in a 48-min training session on the use of TTS in the computer lab with the first author. Students practiced using headphones to hear the audio readings, turning pages with an example story, and logging off when they had completed the reading. Training ended when all students could independently log in, access stories, and use TTS features for one session with 100% accuracy.

Generalization

At the beginning and end of the study, the first author administered the SRI assessment to obtain a current Lexile score to check for generalization of reading comprehension skills. She administered the assessment in the computer lab with all students at one time.

Baseline

The baseline condition (A1) began after students completed their training on the use of TTS. At the beginning and end of this condition, the first author administered an ORF test to measure each student’s words read correctly individually in a separate setting. Throughout this condition, students read the fiction stories on the computer while using only the page-turning icon to progress from page to page. No auditory or highlighting features of TTS were used when reading during A1. There was no explicit instruction on the reading material, nor any specific feedback or reinforcement provided.

When students finished reading, they logged off the computer and answered the reading comprehension CBA questions in pencil-and-paper format without looking back at the text. Students read the stories on the computer without TTS for five sessions in this baseline condition.

TTS as an accommodation

Immediately after students completed baseline sessions, students began using TTS as an accommodation with the stories on the computer in the first accommodation condition (B1). Students accessed the reading in the same way as in A1 but this time enabled the individualized TTS features to provide support while reading. Participants enabled the voice, highlighting, and rate of speed features previously customized from the training session. After reading the story with TTS support, each student logged off the computer and completed the paper-based reading comprehension assessment without TTS support. Students received no explicit instruction on the reading material, no specific feedback or reinforcement, nor any feedback on the use of TTS. Students enabled TTS as an accommodation to support reading for seven sessions. At the end of B1, the first author administered an ORF test to each student in the same manner as baseline.

Second baseline and accommodation conditions

Participants began the second baseline condition (A2) in the computer lab immediately following B1. This condition was conducted in the same manner as A1, without enabling any TTS features, for six sessions. Immediately following A2, students returned to enabling the TTS features on desktop computers to read new fiction stories in the second accommodation condition (B2). These sessions were identical to the B1 condition and continued for seven sessions.

Maintenance

After the second accommodation condition, students moved into the maintenance condition. Maintenance data were collected following the same procedures and instructions as during baseline conditions, on participants’ accuracy for reading comprehension questions. No new instructions were provided to the students. Maintenance sessions were conducted once a week for 4 weeks. An ORF measure was also collected at the end of the maintenance condition.

Interobserver Reliability

Interobserver agreement (IOA) data were collected on the dependent variable during 30% of all sessions or at least once per condition for each student, whichever was greater (Horner et al., 2005). Before the study began, the first author trained an independent observer on how to score the CBAs using similar practice CBAs. Once the observer scored two CBAs with 100% accuracy, she began scoring CBAs from the study. The first author gave a copy of each participant’s CBAs to independently score. The first author and the secondary scorer compared scores after independently grading the CBAs. IOA was calculated by dividing the total number of agreements by the number of agreements and disagreements and then multiplied by 100.

Procedural Reliability

Procedural reliability data were collected in 30% of the sessions for each participant or at least once per condition to measure treatment fidelity. The first author trained the same independent observer on how to record the occurrence of the teacher-directed activities on the researcher-created checklist of procedures prior to the start of the study. The checklist entailed activities in baseline, accommodation, and maintenance conditions (available upon request). The procedural reliability percentages were based on the calculation of the observed components divided by the number of possible components and then multiplied by 100. The resulting calculations determined the mean procedural reliability for each participant.

Results

The primary purpose of this study was to investigate the effectiveness of TTS on the reading comprehension of 9th-grade students with LD. Results showed a functional relation between TTS and reading comprehension performance for all three participants who completed the study (see Figures 1 and 2). Vincent exited the study during B2 due to a change in his school schedule. The effects of the TTS maintained for the three remaining participants. Furthermore, ORF outcomes indicated that all participants increased the number of words read correctly throughout the study (see Table 2). Lexile scores on the district’s universal screener, the SRI, indicated that two of the four participants increased scores between the pre- and posttests.

Figure 1.

Reading comprehension (•) and oral reading fluency (▴) data for Dianna and Donald, respectively.

Figure 2.

Reading comprehension (•) and oral reading fluency (▴) data for Jack and Vincent, respectively.

Table 2.

Comprehension Mean, ORF, and Lexile Scores.

Student	Prebaseline	Baseline	Accommodation	Baseline	Accommodation	Postaccommodation
Student	Prebaseline	A1	B1	A2	B2	Postaccommodation
Dianna
Mean	115	63	83	63	77	77
ORF	772	118	124	119	121	127
Lexile					879	902
Donald
Mean	93	44	70	42	64	68
ORF	972	96	97	95	98	101
Lexile					887	832
Jack
Mean	59	34	64	39	64	75
ORF	186	60	66	66	78	79
Lexile					221	245
Vincent
Mean	110	65	76	62	71
ORF	864	112	116	117	113
Lexile					810

Note. ORF = oral reading fluency.

Dianna

The first graph in Figure 1 represents Dianna’s comprehension and ORF scores during all conditions. Dianna’s comprehension score in baseline (A1) was between 53% and 66% correct with a mean of 63%. A Web-based application, Single Case Research, was used to calculate Tau-U, a measure that estimated the intervention effects by taking into account the nonoverlapping data points and trend between phases (Vannest & Ninci, 2015; Vannest, Parker, Gonen, & Adiguzel, 2016). Tau-U is interpreted as small effect (<.20), moderate effect (.20–.60), large effect (.60–.80), or very large effect (>.80; Vannest & Ninci, 2015). While in B1, Dianna’s accuracy on reading comprehension CBAs demonstrated an immediacy of effect and a Tau-U effect size of 1 CI_90% [.42, 1.00] (p = .00). Her comprehension scores in B1 ranged between 73% and 93% correct with a mean of 83%. After moving from B1 to A2, Dianna’s comprehension score abruptly decreased from 80% to 53%. While in A2, Dianna’s mean comprehension score was the same as in A1 at 63%, displaying repeated lower levels even with one overlapping data point. With the presence of TTS, Dianna’s B2 data were higher than in A2 and the mean for B2 was 77% with a Tau-U of .83 CI_90% [.29, 1.00] (p = .01). Overall, Dianna’s data showed the presence of a functional relation between TTS and reading comprehension with three demonstrations of effect across the conditions. In maintenance, Dianna consistently maintained reading scores with a mean score of 77%.

Dianna’s ORF score started at 115 and ended at 127 by the completion of the study. Using ORF as a progress monitoring tool, Dianna’s ORF in the fall ranged between the 25th percentile (106 WCPM) and the 50th percentile (133 WCPM). Dianna’s ORF at a winter progress monitoring session was 127 WCPM, still in-between the winter progress monitoring of 25th percentile (115 WCPM) to 50th percentile (146 WCPM). The growth expected in that percentile from fall to winter would be 9 in the 25th percentile and 13 in the 50th percentile. Dianna’s WCPM was 12. Hosp, Hosp, and Howell (2007) suggested aiming for the last session when using progress monitoring data. In that case, spring progress monitoring would be in the 50th percentile if the student is below the 50th percentile at the first progress monitoring data session. The suggested growth rate in that time frame, according to Hasbrouck and Tindal (2005), is an additional 9.6 WCPM. Dianna exceeded this with an additional 12 WCPM. Furthermore, her Lexile scores continued to increase throughout the study. Her score was 772 at the start of the study, increased to 879 after B2, and ended at 902 after maintenance.

Donald

Donald’s performance on the comprehension measures during A1 ranged between 33% and 53% on questions answered correctly with a mean of 44% (see Figure 1, Graph 2). An immediate increase occurred when Donald moved from A1 to B1. In B1 with the TTS accommodation, Donald’s accuracy in comprehension questions ranged from 60% to 80% with data moving in a therapeutic, accelerating direction documenting the effectiveness of TTS. The mean of Donald’s B1 phase was 70% and the Tau-U was 1 CI_90% [.42, 1.00] (p = .00). When returning to A2, removing TTS, there was an abrupt decrease to 33%. Donald’s range of answering comprehension questions correctly without TTS was between 33% and 53% and a mean of 42% in A2. Donald’s data demonstrated another immediacy of effect in B2 with only one data point overlapping with data in A2 resulting in a Tau-U of 0.9 CI_90% [.34, 1.00] (p = .01). Donald’s range of answering comprehension questions in B2 was from 47% to 73% with a mean of 64%. Donald’s data showed intrasubject replication through the three demonstrations of effect. In maintenance, Donald’s mean data slightly increased from B2 levels and remained above baseline levels with a mean of 68%. Donald’s overall performance on his CBAs was maintained, further supporting the effectiveness of TTS on reading comprehension outcomes.

Initially, Donald had an ORF score of 93. By the end of the study, his ORF increased to 101. Using ORF as a progress monitoring tool, Donald’s ORF in the fall ranged between the 10th percentile (77 WCPM) and 25th percentile (98 WCPM). Donald’s ORF at a winter progress monitoring session was 101 WCPM, still in-between the winter progress monitoring of 10th percentile (84 WCPM) to 25th percentile (107 WCPM). The growth expected in that percentile from fall to winter would be 7 in the 10th percentile and 9 in the 25th percentile. Donald’s WCPM was 8. Donald did not meet the suggested growth in ORF of 9.6. In addition, he had variable Lexile scores. He went from 872 to 887 after B2 but decreased to 832 at the end of the study.

Jack

Jack’s accuracy on CBAs during A1 baseline was between 26% and 47% with a mean of 34% (see Figure 2, Graph 1). A demonstration of an immediacy of effect occurred when Jack moved from A1 to B1, showing comprehension reading gains with the accommodation of TTS. Once in B1, Jack’s accuracy in reading comprehension scores ranged from 53% to 80% and a mean score of 64% with data moving in an accelerating and therapeutic direction when using TTS. The Tau-U was 1 CI_90% [.42, 1.00] (p = .00). Similar reading comprehension scores were repeated in A2 and B2 with a Tau-U of 1 CI_90% [.45, 1.00] (p = .00), documenting intrasubject replication. Jack’s comprehension scores in A2, without TTS support, ranged from 33% to 47% with a mean of 39% of the comprehension questions correctly answered. In B2 with TTS support, Jack’s comprehension scores ranged between 53% and 73% with a mean of 64%. Jack’s data displayed three demonstrations of effect between A1 and B1, B1 and A2, and A2 and B2. Jack’s maintenance data on reading comprehension CBAs displayed consistent high scores indicating that Jack maintained his reading comprehension skills when using TTS.

Jack began with a fall progress monitoring ORF score of 59 and ended with a winter progress monitoring ORF of 79. This is an increase of 20 WCPM. This nearly doubled the suggested growth rate of 0.6 words per week (Hasbrouck & Tindal, 2005). Jack was far below the 10th percentile with a 59 on ORF measure. In winter, he was closer to a 10th percentile mainly due a growth of 20 WCPM. Jack showed an increase in his Lexile score throughout the study. He went from 186 at the beginning of the study, to 221 after B2, and then to 245 at the end of the study.

Vincent

Vincent’s accuracy on CBAs during A1 was between 53% and 73% with a mean of 65% (see Figure 2, Graph 2). Data within B1 ranged from 53% to 100% with a mean of 76%, and the second half showed data moving in a therapeutic direction. Between conditions analysis for Vincent in A1 and B1 revealed no change and a relative level change of 6% (i.e., 67–73%). His Tau-U was 0.46 CI_90% [−0.12, 1.00] (p = .19) further confirmed no effect. In A2, his comprehension scores ranged between 53% and 67% correct with a mean of 62%. In B2, he had three comprehension scores of 73%, 67%, and 73% correct with a mean of 71% and an increased Tau-U of .83 CI_90% [0.13, 1.00] (p = .05). However, there were few data points in B2 and no data in the maintenance condition due to Vincent changing classes and exiting the study in the middle of B2. These limited data do not show a functional relation like the other participants. Vincent’s ORF score started at 110 and ended at 113 during B2. Because Vincent left the study early, his winter benchmark was not included in progress monitoring data.

Interobserver and Procedural Reliability

Interobserver reliability data were collected on the reading comprehension measure during 30% of the sessions for all participants. IOA was collected on nine sessions (i.e., two in each baseline and accommodation conditions and one in the maintenance condition) for Jack, Dianna, and Donald and seven for Vincent due to his early exit from the study. Jack, Dianna, and Donald’s IOA scores were 100% on all CBAs. Vincent’s IOA was 99% on his CBAs with a range of 96–100%. Procedural reliability data were collected during the same 30% of the sessions for all students as for IOA. The procedural reliability data indicated 96% for Vincent and Jack, 98% for Dianna, and 97% for Donald.

Social Validity

The first author reported data by calculating the participants’ averages in each category of a social validity survey. The participants who completed the survey (Dianna, Donald, and Jack) said they enjoyed using the visual and auditory support of TTS (M = 4.0) on a scale of one to five. The participants were neutral in their response to using TTS’ highlighting and voice selection features (M = 2.7, 3.0, respectively), but they enjoyed using the rate of speed (M = 3.7). With regard to the participant outcomes, participants agreed that they remembered more information after using TTS (M = 3.7); however, they were less likely to use TTS in the future for class assignments (M = 2.7) because they felt it was difficult to access the lab on their own. Lastly, one of the three participants agreed that she or he would use TTS in the future for fun when reading (M = 2.7).

Discussion

The purpose of this study was to investigate the benefit of TTS as an accommodation for students with LD. In addition to seeing the benefits at a curricular level, this study investigated whether TTS would generalize to reading scores on standardize assessments including ORF and Lexile scores. All students who completed this study showed an increase in comprehension of print material using TTS as an accommodation. In particular, students with the lowest reading comprehension scores made the most gains using TTS, while also maintaining their new skills in the maintenance condition. This would be typical of students who started off with the lowest achievement had the most to gain. Below, we will discuss our key findings in relation to each of our research questions.

Reading Comprehension Performance With TTS

In this study, TTS increased the reading comprehension of high school students with LD on curricular assessments. These findings are similar to previous outcomes by Stetter and Hughes (2011) and Stodden, Roberts, Takahashi, Park, and Stodden (2012). Conversely, Gruber et al. (2017) found that only elementary-aged students increased comprehension with TTS. Findings from this study add to the literature base through showing increased outcomes for struggling readers in high school using TTS as an accommodation. When accessing print through a TTS system, students have an opportunity to better comprehend the written text because the TTS releases the cognitive effort required for word recognition, resulting in more effort put toward comprehending reading material. This impacts several areas of instruction beyond reading instruction. At the secondary level, students are frequently required to read traditional print to learn. Social studies, science, and other academic content area present foundational knowledge in textbook form. TTS has proven to be an effective accommodation especially. In the withdrawal phase of our study, it appears that TTS continued to improve reading ability although minimally. However, the improved performance of garnering meaning from the printed text occurred when auditory and visual aspects of the accommodation were used. TTS is a possible support solution for secondary students who are struggling to make meaning from a variety of printed texts.

ORF After TTS

All participants made gains in their ORF from their initial score at the start of the study to their final score in the maintenance condition. In the prebaseline condition, all participants’ ORF scores were well below the 50th percentile. This initial ORF score was used to calculate the rate of speed in the TTS software and was used as the default for future readings. Previous research indicated that when TTS is individualized at a rate faster than readers’ initial ORF, students with LD read more words in less time (Coleman et al., 2011; Young, 2013). Replicating previous research, these students’ ORF continued to increase throughout the study. The second ORF score, after A1, indicated participants’ ORF scores increased one to three words per min. However, at the end of the accommodation condition, only two students increased their ORF by more than the suggested average weekly improvement of 0.6 (Hasbrock & Tindal, 2005), whereas Donald increased his ORF at or just below the suggested level. At the end of the second accommodation condition, three participants increased ORF scores and the participant with the lowest score showed the highest gains.

Social Validity

The survey indicated students generally liked the TTS as an accommodation to support reading. In addition, they generally agreed that TTS helped them to “remember” the stories. However, they felt like they would not use TTS with future assignments because it was difficult to access the computer lab. These results indicated that these students have a very specialized view of the TTS and that TTS is one type of accommodation only offered on lab computers. The students did not appear to perceive that TTS could be offered in a variety of settings or on their own personal computers.

Maintenance of Reading Comprehension Performance

Assessing continued use and benefits of accommodations is an important consideration when planning studies. Few previous studies involved the use of technology and assessed reading comprehension during maintenance (e.g., Kennedy, Deshler, & Lloyd, 2015; Stetter & Hughes, 2011). For example, Stetter and Hughes (2011) found that participants increased their reading comprehension on maintenance probes 2 weeks after exiting the study. In another study, Kennedy, Deshler, and Lloyd (2015) found participants significantly increased accuracy on reading comprehension after three probes in maintenance following the intervention condition. The maintenance data collected over 4 weeks in this study illustrated its uniqueness and addition to the literature. These maintenance data also suggest that when students accessed TTS consistently, they were able to support their comprehension of material better even without the accommodation of TTS, if previously using it. With strong caution, these results could suggest that in addition to TTS being an accommodation, TTS might provide some remediation qualities to struggling readers. However, more in this area would need to be investigated.

Generalization of Reading Comprehension Performance

Overall, participants showed some variability in their Lexile scores throughout the study. Dianna and Jack increased their Lexile scores from pre- to posttest scores by 130 and 59 points, respectively. According to the Lexile growth standards, raising Lexile scores an average of 70 points in one semester is considered average growth (Scholastic, 2007). Dianna doubled more than her average Lexile growth in one semester after using TTS and Jack exhibited average Lexile growth. These results may indicate the potential of generalization to other reading assessments. However, the results should be interpreted cautiously as the teacher could not control for maturation or outside reading instruction in other courses or at home.

Limitations

All studies have a certain amount of limiting factors. The limitations for this study could be considered typical with single subject designs given the small number of participants. When one of the four participants withdrew, issues concerning the generalizability of the findings to a larger population may arise. However, this study does contribute to the larger literature base of AT impacting reading outcomes of secondary students while also replicating results of other studies (Moorman et al., 2010; Young, 2013). Another limitation is that this study occurred in a segregated, self-contained setting. Students learning in a smaller, segregated environment are oftentimes given additional and individualized support as compared to in a general education class. This class had 11 students and 2 adults. Given that the researcher was the classroom teacher and familiar with the students’ needs, her relationship with the students could resulted in her providing student participants with additional unintended support, which could have affected the outcomes.

In addition, this study occurred in the English classroom and not in a reading classroom. All students were enrolled in the same self-contained English and reading classes offering two different curriculum packages during the time of the study. It may be difficult to determine whether the TTS features can be solely responsible for the demonstrated growth. However, the use of a single subject withdrawal design aided in having each participant act as his or her own control. The benefit of using the single subject withdrawal design is to aid in mitigating potential impact of other variables outside of the classroom (e.g., curriculum in a reading classroom) may have on the results.

Finally, given different participant characteristics in this study, it is difficult to conclude that students with the characteristics of LD are the only ones who would benefit from TTS as a valid reading accommodation. Previous research also found that students with comorbid diagnoses (i.e., attention deficit/hyperactivity disorder and LD) may affect reading performance (Wood et al., 2018). Students’ improvements may be related to their secondary labels. For example, Jack’s learning style presented differently from the other students, which raises questions if the results were solely based on the accommodation. Dianna had attention issues which could have affected her outcomes with the use of TTS as an accommodation.

Implications for Practice

Literacy skills are crucial when secondary students are learning content. Technology is a way to alleviate some of the barriers to learning content through reading. Therefore, it is important that students with LD to learn to use technology to increase their acquisition of content. Technology, in particular TTS, is nearly ubiquitous and students should be encouraged to practice using technology to maximize their ability to understand material. Teachers should implement and understand how to teach and use technology to meet students’ needs. Instead, our study indicated that students view TTS as burdensome, specialized equipment only to be used in a lab setting. Using TTS on a regular basis appears to benefit the secondary learner with very few drawbacks noted.

The effects of using TTS could result in long-lasting improvement in understanding content. TTS is sometimes referred to as a “compensatory” tool, indicating teachers no longer need to remediate a reading difficulty rather work around one, removing the barrier of access. However, if students can maintain the improvements made with TTS as an accommodation regularly, perhaps TTS could also be looked at as a remediation tool, not a permanent need.

Our findings imply suggestions for practitioners to consider when planning instruction to increase student outcomes. All teachers should be made aware of the benefits of TTS and the ease at which it can be utilized across content areas for struggling readers. Administrators should promote the use of TTS as an available accommodation for secondary students who struggle with reading by increasing professional development to faculty on the use of TTS when acquiring content. With professional development for teachers and more access to TTS beyond a computer lab, students might improve beyond the one content area. When students can access the meaning of text by being able to comprehend traditional print text using TTS, they will comprehend the print instruction better and therefore achieve at a higher rate and are more likely to engage when using technology. Using technology during instruction for students with LD is a win-win for instructors and learners.

Future research should examine students’ independent use of TTS while exploring ways to increase academic outcomes. More specifically, as technology continues to improve with advanced features, more research is needed to investigate whether other TTS features are helpful accommodations to increase student outcomes in areas other than reading. Also, as students increase their comfort level of using TTS and recognize TTS’s capability of providing access to printed word, they may expand their use of technology to read the printed word. As technology becomes more prevalent in education, using it as an accommodation will assist teachers in customizing lessons to meet their students’ needs, which can result in increased student outcomes.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Anderson-Inman

(2009). Supported eText: Literacy scaffolding for students with disabilities. Journal of Special Education Technology, 24, 1–7.

Assistive Technology Act of. (1998). Pub. L. No. 105-394, 29 U.S.C. § 3002 et seq.

Bouck

Flanagan

Miller

Bassette

(2012). Technology in action. Journal of Special Education Technology, 27, 47–57.

Coleman

M. B.

Carter

Kildare

(2011). How fast is too fast? Determining the speech rate for text-to-speech software. Unpublished manuscript, Department of Theory and Practice in Teacher Education, University of Tennessee, Knoxville, TN.

Cortiella

Horowitz

S. H.

(2014). The State of learning disabilities: Facts, trends, and emerging issues. New York, NY: National Center for Learning Disabilities.

Courtad

C. A.

Bouck

E. C.

(2013). Technology and students with learning disabilities. In Bakken

Obiakor

F.E.

Rotatori

(Eds.), Learning disabilities: Practice concerns and students with LD (Advances in Special Education, Vol. 25, pp. 153–173). Bingley, England: Emerald Group.

Edyburn

D. L.

Higgins

Boone

(2005). Preface. In Edyburn

Higgins

Boone

(Eds.), Handbook of special education technology research and practice (pp. 13–16). Whitefish Bay, WI: Knowledge by Design.

Gast

D. L.

Ledford

L. R.

(2014). Single case research methodology: Application in special education and behavioral science. New York, NY: Routledge.

Gruner

Ostberg

Hedenius

(2017). The compensatory effect of text-to-speech technology on reading comprehension and reading rate in Swedish schoolchildren with reading disability. Journal of Special Education Technology. doi:10.1177/0162643417742898

10.

Hasbrouck

Tindal

(2005). Oral reading fluency: 90 years of measurement (Technical Report No. 33). Eugene, OR: Behavioral Research and Teaching, University of Oregon.

11.

Horner

Carr

Halle

McGee

Odom

Wolery

(2005). The use of single-subject research to identify evidence-based practice in special education. Exceptional Children, 71, 165–180.

12.

Hosp

M. K

Hosp

J. L.

Howell

K. W.

(2007). The ABCs of CBM. New York, NY: Guilford.

13.

Houghton Mifflin Harcourt. (n.d.). READ 180. Boston, MA: Author. Retrieved from https://www.hmhco.com/programs/read-180-universal.

14.

Individuals With Disabilities Education Act Amendments of. (2004). 20 U.S.C. § 1400 et seq.

15.

Kennedy

Deshler

Lloyd

(2015). Effects of multimedia vocabulary instruction on adolescents with learning disabilities. Journal of Learning Disabilities, 48, 22–38. doi:10.1177/0022219413487406

16.

Kurzweil Education. (1996). Kurzweil 3000, Cambium Learning Group. Retrieved from https://www.kurzweiledu.com

17.

Lexile . (n.d.). Registered trademark of MetraMetrics: Durham, NC: MetaMetrics.

18.

Moore

D. W.

Short

D. J.

Smith

M. W.

Tatum

A. W.

(2014). Hampton-Brown edge: Reading, writing, and language, level A. Independence, KY: National Geographic Learning.

19.

Moorman

Boon

R. T.

Keller-Bell

Stagliano

Jeffs

(2010). Effects of text-to-speech on the reading rate and comprehension skills of high school students with specific learning disabilities. Learning Disabilities, 16, 41–49.

20.

Okolo

Diedrich

(2014). Twenty-five years later: How is technology used in the education of students with disabilities? Results of a statewide study. Journal of Special Education Technology, 29, 1–20.

21.

Parette

Peterson-Karlan

(2010). Using assistive technology to support the instructional process of students with disabilities. In Obiakor

F. E.

Bakken

J. P.

(Eds.), Current issues and trends in special education: Research, technology, and teacher preparation (pp. 73–89). Bingley, England: Emerald Group.

22.

Pearson

Inc

. (2012). Aimsweb technical manual. Bloomington, MN: Author.

23.

Perelmutter

McGregor

K. K.

Gordon

K. R.

(2017). Assistive technology interventions for adolescents and adults with learning disabilities: An evidence-based systematic review and meta-analysis. Computers and Education, 114, 139–163.

24.

Poel

E. W.

(2007). Enhancing what students can do. Educational Leadership, 64, 64–66.

25.

Rose

Meyer

Hitchcock

(2011). The universally designed classroom: Accessible curriculum and digital technologies. Cambridge, MA: Harvard Education Press.

26.

Saenz

Fuchs

L. S.

(2002). Examining the reading difficulty of secondary students with learning disabilities: Expository versus narrative text. Remedial and Special Education, 23, 31–41. doi:10.1177/074193250202300105

27.

Scholastic, Inc. (2007). Technical guide. New York, NY: Scholastic.

28.

Smith

Allsopp

(2005). Technology and inservice professional development: Integrating an effective medium to bridge research to practice. In Edyburn

Higgins

Boone

(Eds.), Handbook of special education technology research and practice (pp. 777–792). Whitefish Bay, WI: Knowledge by Design.

29.

Snyder

T. D.

de Brey

Dillow

S. A.

(2016). Digest of Education Statistics 2015 (NCES 2016-014). Washington, DC: National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education.

30.

Snyder

T. D.

Dillow

S. A.

(2013). Digest of Education Statistics 2012 (NCES 2014-015). Washington, DC: National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education.

31.

Stanovich

K. E.

(1986). Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Research Quarterly, 21, 360–407.

32.

Stetter

Hughes

(2011). Computer assisted instruction to promote comprehension in students with learning disabilities. International Journal of Special Education, 26, 88–100.

33.

Stodden

Roberts

Takahashi

Park

Stodden

(2012). Use of text-to-speech software to improve reading skills of high school struggling readers. Procedia Computer Science, 14, 359–362. doi: 10.1016/j.procs.2012.10.041

34.

Strangman

Dalton

(2005). Using technology to support struggling readers: A review of the research. In Edyburn

Higgins

Boone

(Eds.), Handbook of special education technology research and practice (pp. 545–569). Whitefish Bay, WI: Knowledge by Design.

35.

Tian

Okolo

C. M.

(2007). Text-to-speech feature analysis and comparison. East Lansing: Michigan State University. Retrieved from http://rtecexchange.edgateway.net/cs/rtecp/view/rtec-sub/88

36.

U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics, National Assessment of Educational Progress (NAEP). (1992, 1994, 1998, 2002, 2005, 2009, 2013, and 2015). Reading Assessments.

37.

U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics, National Assessment of Educational Progress (NAEP). (2017). Reading Assessment.

38.

Vannest

K. J.

Ninci

(2015). Evaluating intervention effects in single-case research designs. Journal of Counseling and Development, 93, 403–411.

39.

Vannest

K.J.

Parker

R.I.

Gonen

Adiguzel

(2016). Single case research: Web based calculators for SCR analysis. (Version 2.0) [Web-based application]. College Station: Texas A&M University. Retrieved May 22, 2018, http://singlecaseresearch.org

40.

Vaughn

Wanzek

(2014). Intensive interventions in reading for students with reading disabilities: Meaningful impacts. Learning Disabilities Research and Practice, 29, 46–53. doi:10.111/ldrp.12031

41.

Wanzek

Otaiba

Petscher

(2014). Oral reading fluency development for children with emotional disturbance or learning disabilities. Exceptional Children, 80, 187–204.

42.

Wood

S. G.

Moxley

J. H.

Tighe

E. L.

Wagner

R. K.

(2018). Does use of text-to-speech and related read-aloud tools improve reading comprehension for students with reading disabilities? A meta-analysis. Journal of Learning Disabilities, 51, 73–84. doi:10.1177/0022219416688170

43.

Young

(2013). The effects of text-to-speech on high school students’ reading comprehension. Unpublished manuscript.