Abstract
This exploratory study examined the usage of speech recognition (SR) technology by students with high incidence disabilities in grades 4–8 and student and teacher perceptions of using SR as part of the writing process. The study also examined factors contributing to students' use of SR and barriers to using this technology. Results indicated that students across all grades had positive perceptions about using SR, but younger students tended to use it more often. SR was especially helpful for students who struggled with spelling and supported some, but not all, students with drafting text. The study illustrated the importance of taking student variability into account in relation to affinity for SR usage. By integrating opportunities for using SR as part of writing instruction and guiding students to reflect on whether the technology is useful for their individual needs and preferences, teachers can help students with disabilities make choices to use SR in ways that are the most useful for their individual needs.
Writing is a complex task that involves several phases, including planning, drafting, editing and revising. Whereas skilled writers engage in these phases with automaticity and fluency, these aspects of writing can pose challenges for students with high incidence disabilities, such as learning disabilities (LD). For individuals with LD, writing-related challenges include handwriting, organizing, spelling, composing, and editing (Bouck et al., 2015; Graham & Harris, 2013). According to Gillespie and Graham (2014), students with LD typically approach writing as a content generation task. Using this approach, students search their long-term memory for relevant information and generate text by composing each sentence based on the preceding sentence, instead of engaging in a process of brainstorming, organizing ideas, and then drafting sentences with a plan for what they will write. These challenges are compounded when students encounter more complex writing tasks as they progress through the grade levels. Consequently, writing can become a source of frustration and lead to both a lack of confidence and motivation to write and reluctance to practice and master writing skills. Technology tools can be instrumental in supporting students with LD, as well as other students for whom writing poses a challenge, as they learn and practice writing skills.
Speech Recognition as an Assistive Technology for Writing
Assistive technology (AT) tools such as word processing, text-to-speech (TTS), speech recognition (SR), word prediction, electronic dictionaries, and spelling and grammar check have been used for decades to support literacy for students with LD (Bryant & Bryant, 2011). Today, these tools are readily available on mobile devices, no longer requiring specialized devices or expensive software purchases. With many free and low-cost options available, these tools have become more feasible for students to access and use when they undertake the writing process (Bouck et al., 2015; Hunley, 2015). For example, in the prewriting phase, students can use concept mapping tools to brainstorm and organize; in the drafting phase, students can use SR to generate text; and in the editing and revising phase, they can use grammar and spell check software to ensure the technical accuracy of their compositions.
This study focuses on SR, and its potential to support students with high incidence disabilities (primarily LD) during the writing process. In the past, SR software was relatively expensive, and the user had to engage in a process of training the software to accurately transcribe their speaking patterns. Today, SR technology has reached a level of accuracy that makes it easier and more appealing to use to generate speech-to-text quickly (Koch, 2017). There are various SR tools available today ranging from built in features on operating systems (Siri on Apple’s devices), software packages (e.g., Dragon Naturally Speaking), mobile apps (e.g., Dragon Anywhere), and Google Chrome extensions (e.g., Voice Dictation, Google Doc Voice Typing). Generally, the apps and web-based extensions require Internet connectivity whereas the software packages do not.
With increasing access to mobile devices in schools, SR can be a powerful low-cost tool to integrate into writing instruction. SR provides essential supports to address specific challenges that arise for students with LD. With SR, students can (a) generate words more rapidly than typing, (b) bypass challenges of spelling, (c) transcend challenges with short-term memory required to organize and draft text, and (d) express themselves without inhibitions associated with writing accurately or formally.
In a recent literature review of studies of SR in K–12 settings, Pennington et al., (2018) found that existing studies focused on SR’s potential to help users generate text rather than ways in which SR could be integrated instructionally to support the phases of the writing process (e.g., planning, revising, and editing). In addition, studies on SR are dated. Eleven of the 13 studies reviewed were conducted over a decade ago between 1990 and 2010, when SR technology had limitations that no longer exist. For example, SR technology no longer requires training to use, and has become cheaper, faster, and more accurate (Koch, 2017). Only two studies were conducted after 2010, using modern SR tools that include more intuitive user interfaces, better predictive capabilities, faster transcription speeds, greater accuracy, and greater ease of use.
Existing studies have reported positive effects and highlighted the promise of SR to support struggling writers. For example, Higgins and Raskind (1995) found that SR helped students with LD compensate for their difficulties with written composition and allowed students to use their more extensively developed oral vocabularies. MacArthur and Cavalier (2004) found that for students with LD, writing passages generated with SR were of higher quality than handwritten passages. Moreover, SR helped students improve their spelling accuracy (Raskind & Higgins, 1999), generate lengthier passages (McCollum et al., 2014), and improve writing fluency (Garret, 2011). Notably, SR also helped struggling writers improve confidence in their ability to write (Toll, 2014). Over a decade ago, Zhao (2007) predicted that SR would soon be available universally and emphasized the need to systematically develop educational interventions that use the technology in ways that support students’ needs.
Purpose of the Study
This study was conducted at a private school for students with exceptional learning needs. The study focused on students with high incidence disabilities, primarily literacy-related learning disabilities, and was designed to gain an understanding of how students use SR technology that is readily available on their mobile devices. At this school, administrators actively encourage teachers to integrate assistive and instructional technologies in the classroom and provide instructional technology support for teachers and students. As part of the school’s one-to-one computing initiative, all students in grades 4–8 have access to an iPad to use at school and at home. When students started using their devices in class, the school’s instructional technology (IT) specialist noted that using SR features on their iPads had transformed the writing process for some. He made the anecdotal observation that technology was a “gamechanger” for some students who were typically reticent to write; SR appeared to help them persist on writing tasks that were typically challenging. However, not all students seemed to use SR readily even when given access to the tool.
Although existing studies on SR have examined outcomes related to the use of the tool as a support for writing, studies have not examined the student experience of using SR, especially with newer and more advanced SR technologies available today. This study was designed to gain an understanding of how students integrate SR as an assistive tool as part of the writing process. To better understand who benefited from SR and why, this exploratory study examined how and why students made choices to use SR. The study also examined students’ and teachers’ perceptions of factors that facilitated and created challenges related to SR integration with academic tasks, specifically related to writing.
The following research questions guided this study: (a) How do fourth to eighth grade students with high incidence disabilities use SR for dictation on their iPads? (b) What are student and teacher perceptions of the benefits and challenges regarding the use of SR for dictation? (c) What factors and supports contribute to students’ use of SR as part of the writing process?
Method
Participants and Setting
This study was conducted at a private school located in the western U.S. that serves K–8 students with high incidence disabilities and/or students who are gifted. Approximately 70% of the 341 students enrolled in the school were identified with LD or academically at risk and 4% of students were classified as gifted. The school’s mission is to provide a learning environment that supports students with exceptional learning needs. Teachers and administrators at the school are committed to addressing the varied needs of their learners using high-quality instruction and research-based practices in special education, including the use of assistive and instructional technology to support learning. Grades 5–8 are multi-grade classrooms (e.g., fifth to sixth and seventh to eighth classrooms).
Because of the adoption of one-to-one mobile devices across grades 4–8, all students in fourth to eighth grades and their teachers were invited to participate in the study and were provided with consent and assent forms. The majority of participants were identified with LD and some had identifications of attention deficit hyperactivity disorder (ADHD) and autism spectrum disorder (ASD). Participants with no documented disability were enrolled in the school in order to avail of the learning supports available for students with high incidence disabilities. Table 1 presents detailed demographic information on the 95 students and 14 teachers (two 4th grade teachers, eight 5–6th grade teachers, and four 7–8th grade teachers) who consented/assented to participate. All teachers were female and had an average of 13 years of teaching experience (ranging from 1 to 33 years). Nine teachers had bachelor’s degrees and five had master’s degrees.
Participant Demographic Information.
Note. AA = Asian American; AF = African American; ASD = Autism spectrum disorder; C = Caucasian; H = Hispanic; MUL = multi-ethnic; N/A = Not applicable (students without a documented disability identification); NH = Native Hawaiian; OCD = Obsessive Compulsive disorder * Ethnicity info was provided on a voluntary basis and was available for 83 student participants.
Research Design and Procedures
This study utilized an exploratory mixed methods design integrating quantitative and qualitative data (Johnson et al., 2007). The study had two phases: a student survey (n = 95) followed by in-depth interviews of selected students (n = 47) and all teachers (n = 14). The research team included two faculty members at a large public university, the school’s IT specialist, and a doctoral student. The study was approved by the Institutional Review Board (IRB) of the research team’s university. The research team offered an informational session to teachers and students and explained the purpose of the study as part of recruitment for the study. The two phases of the study spanned the course of 5 months. The student surveys were administered in November, data from this first phase was analyzed in December/January, and the student and teacher interviews were conducted in February/March.
SR Instructional Sessions
Because this exploratory study was designed to examine how students use and perceive SR in naturalistic conditions, teachers had flexibility to use the tool for writing-related activities without any specific parameters. However, to ensure that all teachers and students had some foundational knowledge and practice using SR as part of the writing process before the study began, the school IT specialist conducted four 30-minute instructional sessions in all Gr. 4–8 classrooms at the start of the school year, 2 months prior to the first phase of the study. He designed these sessions with the classroom teachers and demonstrated to students how SR can be used as part of the writing process. Each session focused on a stage of the writing process (i.e., planning, drafting, revising, editing). In the sessions, he provided modeling and explicit instruction on how SR could be integrated in each stage. While planning the sessions, the IT specialist worked with teachers to address authentic writing tasks they were doing in class. For example, if students were writing in a daily journal in class, he demonstrated how to use SR for this purpose and gave students the opportunity to practice using SR during journal writing. The instructional sessions included the following activities:
a) Students practiced using SR to brainstorm ideas and plan what to write.
b) Students practiced using SR to draft passages.
c) The IT specialist modeled how to edit and revise sentences generated with SR.
As part of the demonstration, the IT specialist integrated specific tips and strategies for effective use (e.g., how to speak clearly, how to draft and edit later, how to add punctuation manually).
a) The IT specialist showed students how to use SR along with other apps such as Grammarly (grammar checker) and Voice Dream Writer (proofreading app).
The first phase of the study began 2 months after these instructional sessions were conducted, giving teachers and students a chance to incorporate SR into writing activities in class. Because the study was designed to examine SR usage in naturalistic conditions, there was inherent variability between the groups and individuals participating. When this study was conducted, students in grades 5–8 had had access to SR for 2 years. Students in the fourth grade had just received iPads and had been introduced to SR a few months prior to the study. This factor was considered during analysis of data. Teachers were free to use SR as it aligned with their own classroom contexts and teaching styles. Some teachers made it optional for students to use SR whenever they wanted, whereas others used it as an option for specific writing tasks or activities. As a result, the total time, frequency, and tasks for which SR was used varied from classroom to classroom. The variance inherent to the naturalistic approach of this exploratory study was considered during data analysis and related limitations are noted in the Discussion.
Data Sources
Student Survey
A student survey was used to collect data on general usage patterns and affinity for SR. To develop the survey questions, the research team searched previous research on this topic. However, a dearth of studies on user perceptions of SR exist and we found only one related study by Mader (2007) examining the effects of SR technology on the writing skills and attitudes of students with LD. As a part of Mader’s study, participants were interviewed to understand their perceptions about using SR. We developed survey questions based on Mader’s interview questions. In addition, the wording of the items was reviewed by three children at other schools in fourth to eighth grades (who did not participate in this study) to ensure that each item was clear and comprehensible. This 10-item student survey was administered via Google Forms. The items included six close-ended questions and four open-ended questions (see Table 2). Students completed the survey on their iPads in class, taking about 10–15 minutes. The research team analyzed the survey results and identified students who had readily adopted SR and used it often (“high users”) and students who did not use SR much (“low users”). Of the 95 students who took the survey, 47 students, representing high and low users, were invited to participate in follow-up interviews. High users met the following criteria: (a) responded, “yes” to the survey questions Q1 and Q2 (Do you like SR? Do you use SR when you have a choice?) and (b) responded “sometimes or daily” to the survey question Q5 (How often do you use SR?). Low users met the following criteria: (a) responded “no” to survey questions Q1 and Q2, and (b) responded “rarely or never” to survey question Q5. The interview phase was designed to understand in more depth the factors that contributed to students’ preferences for SR usage. Given the exploratory nature of this study, we selected between six and eight participants from each grade level to ensure that our interview data represented a range of perspectives from younger to older users of SR.
Summary of Student Survey.
Note. * For open-ended questions, only responses noted by multiple students were reported in the table due to the limited space.
Student and Teacher Interviews
Student and teacher interviews were used to gain a deeper understanding of perceptions of SR. As was the case for survey questions, the research team developed the student and teacher interview questions based on the study conducted by Mader (2007). We further developed and aligned the student and teacher interview questions to the research questions guiding this study (see Appendix). The student interview protocol included eight questions focused on students’ usage, perceptions, and preferences using SR technology. The teacher interview protocol included seven questions focused on their experiences integrating SR in the classroom and their perceptions of student usage of SR.
Three members of the research team conducted all student and teacher interviews using a semi-structured interview process. Team members posed every question in the interview protocol and asked follow-up or clarification questions as needed. To establish consistency with interview procedures, the researchers paired up initially to conduct two teacher interviews and two student interviews. After ensuring that all interviewers were using consistent procedures in these initial interviews (e.g., initiating the interview in the same way, asking follow-up questions as needed, recording interviews), the remaining interviews were conducted on a one-to-one basis. Student interviews took 10–15 minutes and teacher interviews took 20–30 minutes. All interviews were conducted on the school site (in empty classrooms) and audio recorded for transcription and data analysis. All interview files were transcribed using a professional transcription service.
Data Analysis
Student Survey
Two members of the research team analyzed the survey results and summarized the results independently. For close-ended items, descriptive statistics on frequency of responses were calculated using Excel. For open-ended items, two members of the research team individually hand coded responses and tallied the frequency of phrases or ideas that occurred often. Based on this analysis, they identified common perceptions, outliers, and notable responses. They made notes on areas to explore further during the interview phase. Researchers discussed their analysis of open-ended items and reached consensus on these key findings.
Student and Teacher Interviews
Two members of the research team analyzed interview data, which were audio recorded and transcribed, using qualitative coding methods. First, each researcher examined the interview transcripts individually to identify commonly used phrases, tally the frequency of responses and organize responses into categories related to the research questions. Data from student interviews was categorized into four categories: (a) writing tasks/usage, (b) benefits, (c) challenges, and (d) instruction and support needed. Data from teacher interviews was organized into three categories: (a) benefits, (b) challenges, and (c) instruction and support needed. During a second review of the data, each researcher identified notable responses and highlighted comments that provided further insights related to the research questions. After analyzing the data independently, the researchers met to discuss their analysis and reach agreement on key findings.
Establishing Trustworthiness
Based on measures of trustworthiness for qualitative research (Lincoln & Guba, 1985), the research team used strategies that enhance credibility and confirmability for the qualitative data analysis within this study. Several measures of credibility were used, including data triangulation, investigator triangulation, and member checks. For investigator triangulation, the team worked in dyads to analyze data, and met together periodically to discuss results and confirm findings in each phase. After the initial analysis was completed, the team conducted a member check, providing various means for participating teachers to provide their feedback. All teachers and school administrators were invited to an afterschool meeting where the findings were presented. The research team provided a summary of results for participants to review before the meeting. During the meeting, the team presented key findings and solicited feedback. To ensure that all participants felt comfortable, they were able to give verbal feedback during the meeting as well as submit their written feedback on a collaborative document (Google Doc) during and after the meeting. This allowed participants to respond anonymously instead of publicly in the group setting if they preferred. Participants corroborated the findings and provided related insights that refined aspects of the analysis. To enhance confirmability, the research team maintained an audit trail of research procedures.
Results
Results are reported in the context of the three research questions that guided this study: Student SR Technology Usage (RQ#1), Student and Teacher Perceptions of Benefits and Challenges (RQ#2), and Factors and Supports that Facilitate SR Usage (RQ#3). For the questions that examined perceptions (RQ2 and RQ3), teacher and student data are presented, highlighting commonalities and contrasts.
Student SR Technology Usage
Overall Usage
Student survey and student and teacher interview data were used to examine student SR usage patterns. Teacher interview data revealed that teachers integrated SR technology in varied ways, based on their knowledge of using SR and their personal orientation to integrating technology in the classroom. The ways in which teachers approached SR integration influenced student usage. Based on interview data, some teachers allowed students to use SR anytime; in these cases, students accessed SR for writing tasks as they chose to. Some teachers used SR for specific activities (for example, brainstorming or journaling) and prompted students to use it; in those cases, students used SR only when the teachers initiated the suggestion to use SR. All teachers allowed students to use SR and no one mentioned disallowing SR use for specific tasks.
Table 2 presents results related to student usage of SR.
The following paragraph summarizes responses to questions 1, 4, and 5 on the student survey. A majority of fourth grade students (87%) and over half of fifth to sixth grade students (61%) reported that they would use SR when given a choice (Q1). Students in grades 7–8 were less inclined to choose SR (31%) while 56% of students across all grades reported choosing SR when given a choice. In regard to frequency of use (Q5), 73% of fourth grade students and 60% of fifth to sixth grade students responded they used SR “sometimes”, but most seventh to eighth grade students responded that they “rarely” (44%) or “never” (28%) used it. Across all grades, only a few students (5%) responded that they used SR “daily.” Fifty-three percent of fourth grade students and 43% of fifth to sixth grade students reported they also used SR at home in addition to school (Q4). However, 75% of seventh to eighth grade students responded that they do not use SR outside of school. Only a few students (2%) across grades responded that they used it everywhere. There was a clear pattern across grade levels. Younger students tended to be more willing to use SR and used it more frequently; they also used it across multiple settings compared to the older students.
Writing Tasks
Question 6 on the survey examined the use of SR specifically for dictation tasks related to the writing process. Across all grade levels, students primarily tended to use SR for brainstorming ideas (29%), followed by writing paragraphs for journals, essays, or reports (19%), dictating reminders and notes (13%) and doing school work such as worksheets (5%). Both fourth graders and fifth to sixth graders used SR most often for brainstorming ideas. However, more than half of seventh to eighth grade students (53%) reported they did not use SR for any of the tasks listed (i.e., dictating reminders and notes, brainstorming/organizing ideas, writing paragraphs for journals, essays, or reports, doing school work as worksheets). However, they listed SR for brainstorming as the second most frequent response (22%) followed by “I don’t use SR” (53%). Student and teacher interview data aligned with the survey results. Teachers mentioned that brainstorming ideas was the activity for which students most often used SR. One of the seventh to eighth grade teachers noted,
They [students] use dictation [SR] for brainstorming because they can think and speak faster than they can write. Most of my students cannot type quickly. So, for them, dictation allows them to get a sentence done quicker, so they prefer using dictation than typing.
Corroborating the survey results, student interview data indicated that students often used SR for brainstorming, writing paragraphs (e.g., essay, reports, reading summary), recording information (e.g., taking notes, reminder), and completing assignments (e.g., worksheets). In addition, students reported that they preferred typing to dictating and used SR as a supplementary tool as needed.
Interestingly, both teacher and student interview results indicated that students commonly associated SR with spelling. Even students who rarely used SR noted they used it to spell words accurately. One of the fifth to sixth grade students noted, “I actually don’t use it [SR] a lot but when I do use it, it is for the words that I don’t know how to spell.” Students noted that they used SR when they would otherwise would have had to look up a difficult word. Some students also commented that they did not use SR because they are “good spellers.” The frequency of this response illustrated that many students saw the potential of SR to generate words quickly and accurately but may not have considered how it could be useful beyond spelling a word for them as they wrote.
Student and Teacher Perceptions of Benefits and Challenges of SR Usage
Overall perceptions
According to the student survey results (Q2 and Q3), overall, students tended to have positive perceptions about using SR. Approximately 64% of students responded that they liked using SR (Q2) and 57% of students noted that they think using SR improves their writing (Q3). However, younger students had more positive perceptions about using SR. For example, most fourth grade students (93%) commented that they liked using SR and most of them (87%) said that using SR improves their writing. Fifty percent of seventh to eighth grade students disliked using SR and 66% of them commented SR did not help their writing. Similar to responses about SR usage, there was a clear pattern across grade levels showing that younger students tended to have more positive perceptions about SR compared to older students.
Benefits
Both teachers and students commented that the key benefit was that SR helped students address barriers with writing, especially challenges with spelling words. Teachers used the following phrases to denote the key benefits of using SR: (a) overcome hurdles with writing tasks, (b) write more independently, (c) generate more text, (d) motivate students to write; provides more opportunities to write, and (e) improves student writing confidence. Teachers also noted that SR supported memory and idea generation. For example, a fifth to sixth grade teacher stated, “sometimes written expression can be a real block for them [students]. So [a benefit is] to have an accommodation [SR] that can really help them with that.” Another 5-sixth grade teacher stated that using SR could help students “free their space so they can really get their thoughts, their genuine thoughts, down without the hindrance of having the actual act of writing in the way.”
Teachers commented that SR helped students write more quickly, with more ease, and allowed ideas to flow. Using SR, students were able to find the spelling and meaning of difficult vocabulary independently. For example, a fifth to sixth grade teacher remarked that students might quickly ask Siri to spell words like amoeba or paramecium instead of asking the teacher for spelling assistance. This teacher also noted, “They [students] do it [writing] more often without resistance because they don’t hate it so much now. It is more jovial and less pushing them up a mountain to get them to do the assignment.” A seventh to eighth grade teacher also commented that her students generated “two, three, or four times as many sentences” with SR. In these cases, teachers reported that SR helped students draft text more quickly, allowing them to speak and capture what they were thinking, and ultimately improved their confidence with writing.
Students used the following phrases to denote key benefits of using SR: (a) overcome writing challenges, (b) spelling support, (c) easy to use, and (d) write faster than typing. Some students commented on their use of SR for spelling. A fifth to sixth grade student stated, “I like it [SR] because I suck at spelling. I basically cannot spell, so it helps me spell correctly.” Some students also noted that being able to spell accurately via SR helped them to write independently and improved their spelling skills. Students also mentioned that they could generate text quickly with SR and it helped them to “save brainpower” needed for writing, allowing them to focus on expressing their ideas. A fifth to sixth grade student noted, “It opens a whole new learning for all these kids…If people have trouble with writing or typing, they can express it in their own talking way…it is just amazing.” Students also noted SR was easy to use and could benefit all people, including students who struggle with writing. A seventh to eighth grade student said, “I use it [SR] for writing, other people use it for research. There are many uses for dictation [SR] and I think that can help everybody in some way.” Finally, students also noted the mobile aspect as a benefit; they could use SR at anytime and anywhere because it was a feature of their iPads.
Challenges
Based on both the surveys and interviews, inaccurate transcription was the biggest challenge of using SR, according to students. In the survey (Q8), 67% of students across all grades responded that inaccuracy was a challenge of using SR. During interviews, students commented often on the inaccuracy issue, stating that SR did not understand certain words such as proper nouns or last names. Also, SR did not accurately transcribe the speech of students with non-standard accents and enunciation difficulties (e.g., speech impairments) and inaccuracies increased in loud classrooms settings. A fifth to sixth grade student noted, “Because it [SR] was inaccurate, I told the teacher that I didn’t want to use it.” Students who were reluctant to use SR often commented on inaccuracy as a barrier. Students tended to attribute the inaccuracy to the technology itself. However, student and teacher interview data revealed that inaccuracy arose at times due to the user or environment, rather than the technology. For example, some inaccuracies were due to articulation challenges, students speaking too fast or not enunciating clearly, and/or using SR in a loud environment. Students noted that at times the Internet speeds slowed down and created a lag between speaking and having the words show up on screen; they found this delay frustrating as they were trying to dictate.
Teachers commented on various challenges including the ones identified by students. However, there was no challenge that stood out among teachers as the greatest one. The following are various challenges noted by teachers: (a) difficult to find the right environment (e.g., quiet place), (b) student distraction and improper use, (c) lag of typing due to Internet connection issues, (d) inaccurate transcription, (e) students’ expressive language difficulties, (f) students’ resistance and anxiety of using SR, (g) students’ editing skills, and (h) limited teacher knowledge and experience. First, teachers noted it is important to find the right environment for SR use to promote accurate transcription, and that it could be challenging to find a quiet place within the classroom. Some teachers allowed students to go outside the classroom and find a quiet place when using SR, but some did not feel comfortable letting students work unattended. A fifth to sixth grade teacher stated that finding the right environment for SR to work well for students was “the most frustrating part.”
Second, teachers across all grades noted that students became distracted at times when using SR on their iPads. One of the seventh to eighth grade teachers noted that students had free access to “check their text and e-mails and update their Instagram.” One of the fifth to sixth grade teachers also noted that students became distracted by humorous or ridiculous errors that resulted from inaccurate transcription. A seventh to eighth grade teacher noted, “they [students] fixate on it [errors] and they want to show all friends.” One of the fifth to sixth grade teachers also commented, “Many of students just fool around with it [SR]. They don’t use it properly.” Third, teachers also described time lags when the Internet connection was slow and noted that it created frustration for students. Some teachers mentioned that students complained that Internet was required for the SR app to work. (Mobile-based SR apps, such as Siri, use an Internet connection for the transcription, making it impossible to use SR in a place without Internet.)
Fourth, teachers also commented about the issue of inaccurate transcription. However, they did not seem to find it as serious a drawback as students had. One of the seventh to eighth grade teachers noted, “they’ve [students] complained about that [inaccuracy]…I found with middle school, they’ve become very fixated on when it [SR] either autocorrects incorrectly or it doesn’t pick up the correct word.” A seventh to eighth grade teacher commented that inaccuracies with place names and proper nouns were a challenge and her response was to encourage students to go back and edit those manually. Fifth, teachers noted that students who have expressive language difficulties or non-standard accents can become frustrated with SR. One of the fifth to sixth grade teachers noted that for one of her students who did not have a typical speech pattern SR affected self-esteem negatively, and “she became frustrated and discouraged. The difficult part was getting her to persevere…she kind of looked at it as a burden.”
Finally, teachers commented about students’ resistance to use SR even though it would help them draft text. One of the seventh to eighth grade teachers noted she was “surprised at their resistance” and she thought, “They [students] haven’t given it [SR] the chance.” Two fifth to sixth grade teachers commented that students have “a lot of anxiety and…they don’t want people to hear them dictating.” One of the seventh to eighth grade teachers also noted that students felt “embarrassed to speak out loud” when using SR. Seventh, teachers noted that students did not want to go back and edit after generating text with SR. Eager to get the writing task over with, they felt as though the text generated by SR should be the final draft and resisted editing it. One of the seventh to eighth grade teachers noted that going back to edit and revise what they drafted with SR would be “another hurdle” for the students. Last, a couple of teachers also commented about their own limited knowledge and experience as a challenge. One of the seventh to eighth grade teachers commented, “I am limited in my experience of doing the app [SR] that I may not know how to help the students. That’s why I said I think I need more training.”
Factors and Supports That Facilitate SR Usage
Usage factors
To determine whether there were specific factors that predisposed students to use SR or not, we examined the responses of the 47 students who participated in both the surveys and interviews. First, high users tended to have more spelling difficulties than low users. High users tended to report they had spelling problems due to their disability and felt SR helped their writing because it helped them write faster. Low users tended to report they were good spellers and had no need to use SR. They also tended to think that SR did not help their writing. Second, high users tended to have strong oral skills based on information provided by their teachers during the teacher interviews. Teachers and students noted that students with expressive language difficulties (e.g., speech impairment, accent) struggled with using SR. According to teachers, students who had severe spelling problems and proficient oral skills tended to be more willing to use SR.
Support needs
To examine the support needs required to facilitate the use of SR, we examined student survey data (Q9) and interview data. According to the survey data, students across all grade levels reported that increased accuracy of SR technology would make them more likely to use it. Students also reported that no lag between dictation and text generation and no need for Internet would also facilitate their use of SR. During interviews, students also reported that “a quiet place” and “use of a headphone with a microphone” would be factors that predisposed them to use SR more frequently. One fifth to sixth grade student noted:
Sometimes, the class could be a little loud for dictation. That’s why sometimes it will catch other things and it will add more stuff that I don’t need. I usually go outside and dictate outside because it’s quiet.
One seventh to eighth grade student noted, “Being in a quieter place. Using a microphone or headphones. And also enunciating…. it would be able to understand what you are trying to say.”
Students also recalled suggestions and tips that the IT specialist had provided and demonstrated at the SR orientation sessions a few months prior to the start of this study. Tips that they remembered most often were (a) speak clearly, (b) use other AT tools along with SR, and (c) do not worry about editing as you dictate; go back later to edit the text. In regard to using other AT apps that they were familiar with, one of the fifth to sixth grade students noted, “You can start using Grammarly after to get what SR spelled wrong.” He mentioned that using this grammar checker helped him identify and edit inaccuracies in passages he had dictated using SR.
Teachers commented on supports that could facilitate their integration of SR with instruction. The support needs were organized into three categories: environmental support, device support, and instructional support. First, environmental support noted by teachers included (a) need for a quiet place, (b) comfortable environment for speaking out loud, and (c) better Internet connection. One of the seventh to eighth grade teachers described the difficulties in allowing students to go outside to find quieter places. “Sometimes going to hallways or outside to find quieter places is distracting for them and it is not necessarily quieter.” She pointed out that while it is important to find a quiet place, it is also important to be able to monitor students when they are outside the classroom. In addition, one of the seventh to eighth grade teachers mentioned “the first obstacle that I’ve seen is a socioemotional one…. in terms of making the environment comfortable.” She felt it would be helpful for teachers to emphasize that SR is an appropriate assistive tool to use for writing and to help reduce the shyness some students feel about speaking aloud with others around.
In terms of device support, teachers mentioned that the following would help: (a) a headphone with a microphone and (b) improved SR accuracy. One fifth to sixth grade teacher noted that a headphone with a microphone “was easier especially for those who don’t have such clear speech patterns.” Teachers reported that a microphone may increase accuracy, especially for students who have expressive language difficulties.
Last, instructional support included (a) having students practice verbal skills, (b) teaching students how to edit, (c) providing more support and scaffolds for writing and editing (e.g., guided prompts for writing, structured checklist for editing), (d) providing more opportunities to use SR, (e) using SR with other AT apps, and (f) providing more SR training for both teachers and students. Teachers noted that students could benefit from understanding how to clearly articulate words to enhance accuracy, for example, speaking more slowly and clearly (e.g., breaking words down into syllables) and avoiding utterances such as “um” or “well.” One of the fifth to sixth grade teachers noted students’ reluctance to edit saying that they should “learn literacy-related things before using SR…so SR is another layer where they need to [have] other deficits met.” One teacher noted that students would be more likely to revise passages generated through SR if they learned editing strategies in conjunction with SR usage. She suggested incorporating mnemonics such as COPS (Capitalization, Organization, Punctuation, Spelling) or providing prompts or checklists.
Teachers also noted students should have more opportunities to use SR to see its value. Some teachers expressed concerns that students do not understand its value and just “fool around with it.” A fifth to sixth grade teacher noted, “The SR technology is vastly helpful to them. It is not like they are not ready for it.” One of the fourth grade teachers suggested that teachers should “provide more projects that will encourage them to use dictation.” If students are given opportunities to use SR, they can see and experience the potential of this tool as part of the writing process. Finally, teachers reported the training provided by the school IT specialist was very helpful for both themselves and their students. Some teachers noted they were still not comfortable with providing support with SR and would appreciate additional training. They mentioned the need for support as they integrated SR as part of writing instruction. Teachers noted that explicit instruction on using SR for the writing process could help students understand how to generate and edit text. Students would also benefit from opportunities to practice these skills. Two teachers mentioned the possibility of including scaffolds for writing along with SR. For example, one idea was to provide prompts, allowing students to generate responses by SR and then to reorganize the text they generated into written passages.
Discussion
This study examined factors that influence student use of SR and highlighted ways in which personal preferences, technical challenges (perceived and real), and specific needs and characteristics of the individual influenced student use of SR. Based on the results of the study, we present considerations and implications for using SR as an assistive and instructional technology support for students with disabilities.
Instructional Strategies That Enhance SR Usage
SR technology is simple and intuitive to use; when invoked with a click of a button or by voice input (‘hey Siri’), SR immediately begins converting speech to text. Although using the technology is simple, there are additional considerations to ensure that it is used optimally during the writing process. Higgins and Raskind (2000) suggested providing instruction on SR for writing. Students can benefit when teachers model how to use SR, making explicit how SR can be integrated during each phase of the writing process. The benefits of instruction on how to use SR in specific ways were illustrated in this study. The IT specialist had provided explicit instruction on each phase of writing and given students opportunities to practice various writing tasks, such as brainstorming and drafting text with SR during his training sessions. During interviews that occurred 5 months later, several students recalled and referenced these tips. With the modeling and opportunities to practice using SR as part of the writing process, students may develop a better understanding of when and how to use SR to support them with their writing tasks. Researchers have noted students with LD can improve their writing with SR with explicit instruction (MacArthur, 1998; MacArthur & Cavalier, 2004). As students learn how to integrate SR into the writing process, they will become more aware of how technology can support writing and address the challenges they face, such as with spelling and generating text based on their ideas. For students with LD, it can also be beneficial to teach and model how to use other AT apps, such as grammar checkers (e.g., Ginger or Grammarly) and literacy support apps (e.g., Voice Dream Writer) along with SR. By bringing together a suite of AT tools available on their devices, students can fully avail themselves of the tools that address the challenges they might typically face when writing.
Students can also benefit when they can make choices to use SR. Teachers can introduce SR as they teach foundational writing procedures and strategies (MacArthur, 2009) and create opportunities for students to try using SR. By having the opportunity to use SR and reflecting on whether it is useful for them, students with disabilities can be more self-directed about using this AT tool. SR may also not be appropriate for all students with high incidence disabilities; some students simply prefer typing and can generate text more quickly and accurately by typing instead of speaking.
As this study indicates, SR is more useful for some students than others; having opportunities to try SR within authentic writing tasks will allow all students to gauge its usefulness for themselves. Optimally, SR should be a choice for students to use for writing anytime, and not limited to specific tasks or activities. Graham and his colleagues (1998) noted that when students realize the value of SR, it can also increase their motivation to write.
Influence of Age and Grade Level on SR Usage
Although students across grade levels had positive perceptions about using SR overall, younger students in this study tended to be more positive and used it more often than older students. One reason for this might be that younger students are still developing their writing skills and not yet set in their ways. Consequently, younger students may have been able to assimilate SR into their writing process more easily. Older students have been writing for several years and had learned strategies to address and accommodate challenges with spelling, grammar, and content. Due to their prior experiences, SR may not have been as natural to integrate because they are used to typing and using other accommodations and modifications for writing. Older students might also have held preconceived notions about what tools were acceptable supports. One seventh grader mentioned that using SR was like “cheating.” Although students at this school regularly use technology as an instructional support and are encouraged by teachers to avail of assistive technology tools (e.g., TTS for reading), this student did not view SR to be a legitimate assistive tool.
In this study, the younger students (fourth graders) expressed more affinity for SR than their counterparts in grades 5–8. The novelty factor and attitudes related to their age and experiences with learning how to write may have influenced this. The fourth graders had started using iPads and SR earlier in the year while older students had used these tools for several years. These younger students expressed less shyness about speaking aloud to dictate and approached new tasks with less resistance and more curiosity. Some of the older students, who were pre-teens and teenagers, felt more embarrassment speaking aloud. Younger students’ writing skills were at the emergent phase; they wrote shorter passages and were still learning grammar and the revision process. Older students had greater writing demands in terms of length, grammatical accuracy, and organization and therefore might have found it harder to integrate a new tool into a process they had been learning for some years.
Implications for Practice
Findings of this study aligned with previous studies (e.g., Higgins & Raskind, 2000; MacArthur & Cavalier, 2004) reporting that SR can remediate writing deficits of students with LD. These prior studies found that SR effectively helped students with LD improve writing skills, such as spelling and writing quality. According to researchers, students generated longer and more complex texts and were able to write significantly faster when using SR compared to typing or handwriting (De La Paz, 1999; Graham, 1999). MacArthur (1999, 2009) noted SR could cause a “cognitive burden” for elementary to high school students; for example, students must articulate clearly and learn how to correct spelling errors. Although modern SR technology has greater accuracy than it did 10–20 years ago, this study demonstrated that inaccurate transcription still poses a barrier for students, creating a cognitive burden while writing for some.
Based on prior research and the findings of this study, we present the following considerations that can be useful for teachers, teacher educators, and AT/IT specialists seeking to use SR for students with high incidence disabilities. First, while SR itself is easy to use, students can benefit from instruction on how to embed SR into each writing stage. Teachers can model, provide guidance, and opportunities to practice using SR in each phase of the writing process. It is useful to emphasize the need to review and edit text generated with SR. This can set the expectation that initial drafts of writing, whether they are handwritten, typed, or generated with SR, should be reviewed and edited. Second, teachers should give students opportunities to practice using SR and reflect on whether and how it helps them. This can help students become more self-aware of when they might want to use SR. Third, teachers should teach students how to use other literacy-support AT apps along with SR as they undertake the phases of the writing process. Fourth, students can benefit from having quiet spaces when they are using SR. Providing students with a headphone and microphone can also help to enhance accuracy. Fifth, students with severe spelling problems perceived SR as a support to meet their needs, whereas other students with LD did not find it as useful. Forgrave (2002) notes that SR is an effective tool, especially for students whose oral communication skills are better than their writing skills. Students can benefit from demonstration and practice with SR, allowing them to see how it can support their writing process beyond the spelling feature. Finally, inaccurate transcription was a challenge identified by students and teachers. Inaccuracy can result from improper use of SR (e.g., speaking too fast, loud environments) and challenges related to articulation. Teachers should provide explicit instruction on how to pace one’s speech, how to articulate clearly, and how to edit passages after generating text in order to help students feel less inhibited by inaccurate transcription.
Limitations and Future Research
There are several limitations to this study. The results of this exploratory study are limited to the context of one school setting and are not generalizable to other students and different settings. The study was conducted in a private school serving exceptional students, which created some unique factors. Teachers at this school receive technology training and support and have a high comfort level with the use of technology as an instructional support for students. As a result, they regularly allow and encourage students to use assistive and instructional technology tools in the classroom. In other school settings (e.g., public schools), technology usage and experiences may be quite different if teachers do not have the training or comfort level to integrate technology into instruction. In addition, all students had access to the same type of device (iPads), making it easier for teachers to integrate SR into instruction. In other settings, it may not be possible to ensure that all students have the same type of device, which can impact student and teacher usage and perceptions.
Because of the naturalistic and qualitative nature of this study, we did not limit participation to students with specific characteristics (e.g., disability categories, length of exposure to SR). All students in fourth to eighth grade were invited to participate in the study. Due to school data privacy policies, we could not collect data that allowed us to make a correspondence between student characteristics (e.g., disability categories, vocabulary skills) and their usage patterns. This poses a limitation to the study because we are unable to draw conclusions about usage patterns in relation to specific student characteristics.
Our purpose was to better understand SR usage in naturalistic conditions, so controlled conditions (e.g., duration or frequency of usage) were not introduced. Instead, this study relied on student and teacher perceptions and self-reporting of SR usage. Since external measures of performance were not collected (e.g., vocabulary achievement scores, student SR usage patterns), this study is limited in its scope. The results of this study provide a foundation to further examine how teachers can enhance SR use to support students.
Future research on SR is required to broaden our knowledge base of the potential of integrating SR with instruction and the outcomes and efficacy of SR use. This study examined SR in a school where assistive technology tools are actively used by teachers in all classrooms. Future research should examine how SR can be integrated in varied settings (e.g., public schools, inclusive classrooms, resource rooms) and grade levels (e.g., secondary school students who may benefit from SR for the advanced writing tasks required in high school). When examining stakeholder perceptions and usage, researchers can use technology self-efficacy measurement tools (e.g., Holcomb, King, & Brown, 2004) to determine attitude, knowledge, skills, and use of technology and include classroom observation to enhance the credibility of self-report data and to gain a deeper understanding of when and how teachers integrate SR into instructional task. It will also be useful to examine how teachers’ characteristics (e.g., teaching experience, knowledge and perceptions about SR, instructional styles) and students’ characteristics (e.g., disability categories, vocabulary skills) can affect students’ SR usage.
Future studies that use experimental research designs and rigorous research methods will help to establish the efficacy of SR on student writing outcomes, especially in relation to instructional strategies that can be used in conjunction with SR. For example, researchers can examine more targeted strategy instruction on using SR, by providing procedural knowledge about SR (how to use the tool), followed by conditional knowledge (when to use SR), and then by process knowledge (how to generalize or modify SR for other purposes). This will allow researchers to make determinations about how students apply and generalize the use of SR and consider for whom strategy instruction on SR is most beneficial.
Conclusion
Readily available on our mobile devices and laptops, SR is a useful technology support for students with and without disabilities. For students with disabilities who struggle with writing tasks, SR can be instrumental in overcoming barriers. However, it is important to consider individual student variability when it comes to affinity for SR usage. The results of this study illustrated that some students with disabilities benefited from and used SR more than others. Ideally, students should understand when and how SR supports their writing process and be able to make informed choices to use the tool as needed. As educators, we can integrate opportunities for using SR as part of writing instruction and guide students to reflect on whether the technology is useful for their individual needs and preferences, thereby giving students with disabilities the agency to make choices and use SR in ways that are most useful to them. Ultimately, this will allow students with disabilities to make self-determined choices about when and how to use this assistive and instructional technology tool to support their writing process now and in the future.
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.
