Assistive Technology Use by Students With LD in Postsecondary Education

Overview of Delivery of AT Services at Postsecondary Institutions

As previously stated, the provision of AT in postsecondary institutions is organized, in most instances, by DSOs. According to Raskind and Higgins (1998) DSOs employ either a distributive location model whereby AT can be accessed throughout the postsecondary environment or a central location model where AT is used primarily at one site. These two approaches can be viewed as extensions of the two main models of special education delivery used within elementary and secondary school for years: mainstreaming/integration versus withdrawal/segregation. The distributive model tends to facilitate integration of students with LD into the mainstream student body and is consistent with the principles of universal instructional design, a dominant pedagogical format in postsecondary environments, which calls for multiple means of representation, action and expression, and engagement. However, the need for increased funding to effectively deliver this model and a lack of institutional resources (e.g., specific AT knowledge at different locations, space, staffing, and hardware requirements) have been found to be major impediments to the implementation of this model (Abreu-Ellis & Ellis, 2006). The central location model has been found to result in greater levels of student satisfaction and success as well as more efficient delivery of services when compared to a distributive model (Burgstahler, 1992). However, it has been criticized for creating stigma in association with the use of AT through its sole and thus obvious point of distribution. Perhaps this is why the majority of postsecondary disability service providers in the United States (Burgstahler, 1992) and Ontario (Abreu-Ellis & Ellis, 2006) endorse the distributive location model of AT.

Disability service offices at Canadian postsecondary institutions experience some common challenges in relation to providing specific AT services to students with LD. One core challenge, according to Abreu-Ellis and Ellis (2006) involves fitting AT to the needs of a given individual as university disability service providers have limited resources and thus a limited range of products at their disposal. This unfortunately can lead to “cookie cutter solutions” in which the student must try to fit the solution rather than the student being offered a solution tailored to their needs. Another challenge faced by postsecondary disability service providers pertains to the background expertise of personnel providing training in AT and the number of personnel per office available to provide AT service. One fourth of universities in Ontario have stated that they do not have the personnel resources required to meet the AT demands of students with LD (Abreu-Ellis & Ellis, 2006). The DSOs in Ontario colleges are unevenly staffed with some having staff dedicated to a role such as Assistive Technologist or Learning Strategist, or Disability Advisor, while other offices have a single staff member working to offer all three services (College Committee on Disability Issues, 2009).

A progressive initiative in the province of Ontario to develop consistency in AT service delivery at postsecondary institutions was achieved through the creation of the Learning Opportunities Task Force (LOTF) in 1997. The LOTF was mandated to (a) improve the transition of students with specific learning disabilities from secondary to postsecondary school and (b) enhance the services and supports that students with LDs receive within the postsecondary sector (LOTF, 2002). The LOTF recommendations resulted in Ontario postsecondary institutions implementing higher quality and more uniform AT services for postsecondary students with LD in comparison to pre-LOTF services. For example, learning strategists and assistive technologists were mandated for all postsecondary institutions in Ontario. More specifically, assistive technologists were given the role of providing the technical expertise and coaching that LD postsecondary students required to independently employ AT within their coursework or program of study. In addition, to help address the variations in expertise of AT personnel at disability service offices LOTF recommendations facilitated a province wide online AT certification program for assistive technologists.

Overview of Assistive Technology Trends at Postsecondary Institutions

AT use occurring on U.S. and Canadian campuses presents as quite similar and is outlined in Table 1. Research conducted across Ontario universities determined that ATs related to reading and writing were among the most commonly accessed by students with LD (Abreu-Ellis & Ellis, 2006). This included software capable of text to speech, word prediction, voice recognition, and brainstorming/mind mapping software. The available U.S. research on AT usage was completed using a composite measure of students with all types of disabilities who accessed disability service centres; again ATs designed to support reading and writing were among the most frequently used (Sharpe, Johnson, Izzo, & Murray, 2005). A sole exception indicated within the Canadian data was brainstorming/mind mapping software (to assist with writing), which was used frequently by students in Canadian PSEs. Although the research investigating trends in AT use at PSEs is scant, to say the least, it is consistent in both U.S. and Canadian samples and it coincides with research indicating that these academic skills are the most critical for success in PSE (Martinez-Marrero & Estrada-Hernandez, 2008; Mull & Sitlington, 2003).

OPEN IN VIEWER

Table 1.

Comparison of Most Frequently Used Assistive Technology on Canadian and American Postsecondary Campuses.

Rank order (descending order of use)	Canadian postsecondary institutions a	American postsecondary institutions b
1	Text-to-speech (Kurzweil™)	Audio books
2	Voice recognition (Dragon Naturally Speaking™)	Portable word processor
3	Outlining/Graphic mind mapping (Inspiration™)	Word prediction with speech output
4	Word prediction with speech output (Texthelp™)	Text-to-speech
5	Screen reader (without OCR) (Read Please™)	Voice recognition
6	Portable word processor (Quick Pad™)	Screen reader (without OCR)

a.

Rankings of learning-disabled students’ assistive technology use by disability service representatives from 17 universities in Ontario (Abreu-Ellis & Ellis, 2006).

b.

Rankings of assistive technology by students of all disability categories at American 2- and 4-year postsecondary institutions (Sharpe, Johnson, Izzo, & Murray, 2005).

Description of the Most Frequently Used Postsecondary Assistive Technology for Reading

Assistive Technology for Writing

Literature Review of the Efficacy of Assistive Technology to Remediate Postsecondary Academic Deficits

Despite the increasing use of AT by postsecondary students with LD there is a paucity of empirical research examining its efficacy to circumvent academic deficits. Indications of the effectiveness of AT in PSE have been derived primarily from student testimonials, anecdotal reports, and case studies (Fichten et al., 2003; Raskin & Higgins, 1995). Similarly, a review of the efficacy of AT studies conducted with children of all manner of disabilities (Watson, Ito, Smith, & Andersen, 2010) characterized the research as primarily qualitative, single subject, and nonexperimental. More rigorous research designs are clearly needed to investigate the efficacy of AT in both child and adult populations. What does exist within the literature, however, that holds some relevance to the issue at hand is some methodologically sound studies examining AT as a remedial tool for children with LD (see National Reading Panel, 2000; Wanzek et al., 2006). The findings from these studies can broadly be summarized as supporting the notion that AT is effective as a remedial tool for reading and spelling deficits. In addition, the National Reading Panel (2000) has suggested that text-to-speech AT is a promising instructional technique for reading instruction. Although this lends some credence to the idea that AT may therefore go on to hold a role as a tool capable of circumventing the academic deficits of children and possibly even adults with LD, this notion must be viewed as one created from extrapolated and incomplete evidence.

A small number of studies in the early 1990s investigated the use of word processing on academic outcomes for postsecondary students with LD. The results of this research indicated that the use of word processing helped students complete courses at a rate comparable to nondisabled peers, improved GPA, and facilitated detection of spelling errors (Collins, 1990; NcNaughton, Hughes, & Clarke, 1993; Primus, 1990).

Research by Elkind, Black, and Murray (1996) examining the efficacy of AT for reading examined the role of text-to-speech software on the reading performance of adults with learning disabilities. The subjects in this research study consisted of university and college students and adults who were in the workforce or seeking employment. LD inclusionary criteria for the study consisted of a discrepancy-based definition of LD (IQ vs. achievement) endorsed at the time by the National Joint Commission on Learning Disabilities (Hammill, Leigh, McNutt, & Larsen, 1981). The researchers reported that the group of dyslexic students using text-to-speech showed enhanced reading rate, comprehension, and ability to sustain attention to reading compared to the dyslexic group who did not employ this software. Higgins and Raskind (1997) also found reading comprehension scores improved for a sample of severely reading disabled postsecondary students when they used a text-to-speech system. However, an inverse correlation between unaided reading comprehension scores and reading with text-to-speech was also noted when the sample was divided into two groups based on the degree of reading impairment (i.e., severely impaired and less impaired reading comprehension groups). This led the researchers to qualify their findings and assert that text-to-speech systems assisted some students greatly but seemed to interfere with the performance of others. In essence, the text-to-speech software appeared to interfere with the reading comprehension of the less impaired readers yet it enhanced the comprehension of severely impaired readers as evidenced by the change in their scores during the condition of unaided reading comprehension.

Higgins and Raskind (2005), summarizing the research on AT for use with reading (i.e., primarily text-to-speech software), commented that the evidence to date indicates that text-to-speech systems seem capable of both helping and hindering the reading comprehension of students with LD. This conclusion is based partially on their earlier work (Higgins & Raskind, 1997) showing that the greater the reading disability, the more likely the technology was to elevate reading comprehension scores, and conversely the milder the reading disability, the more technology seemed to interfere with reading comprehension. The researchers hypothesized text-to-speech systems that read every word aloud might interfere with reading comprehension by overtaxing working memory. A further study by Higgins and Raskind (2005) provided some support for this hypothesis as they found the reading comprehension scores of adolescents with LD (mean = 14.2 years) improved through the use of a handheld text-to-speech scanning device designed to read only selected, single words. To explain this finding, Higgins and Raskind (2005) stated that continuous auditory input whereby every word is read aloud by text-to-speech programs have more potential to disrupt working memory processes related to reading comprehension than single word text-to-voice programs. It is important to note that although Higgins and Raskind (2005) speculate that working memory processes are implicated in the effectiveness of AT on reading comprehension, this explanation is not based in data as they did not measure or assess working memory abilities in their study.

More recently, Holmes and Silvestri (2009), using a small sample of adults with self-reported reading difficulties, found that text-to-speech software improved the reading comprehension scores of individuals with specific deficits in phonological processing. Conversely, the reading comprehension scores of adults with a different profile of processing limitations (i.e., average or better phonological processing and rapid naming abilities) were hindered by this technology. The researchers evoked Sweller’s cognitive load theory (1994) to account for this interaction between phonological skills, reading comprehension, and text-to-speech software. More specifically, it was hypothesized that the adults in this sample who benefited from text-to-speech (i.e., those with poor phonological processing skills) did so because the instructional format (continuous auditory input) and intrinsic processing load (compensation for processing skills deficits) created a germane cognitive load leading to gains in reading comprehension. Conversely, subjects with intact phonological processing skills did not display reading comprehension gains when using text-to-speech software, as the instructional format (every word read aloud) and intrinsic processing load (interference with processing skills) produced a detrimental cognitive load and, ultimately, poorer reading comprehension performance ensued.

Raskind and Higgins (1995) also examined the effects of text-to-speech software on the proofreading performance of postsecondary students with LD. The results indicated that text-to-speech synchronized with highlighted text on the monitor allowed students to identify a higher number of writing errors in comparison to using a human reader or a no assistance condition.

A literature review was unsuccessful in locating peer-reviewed studies related to the use of speech/voice recognition, word prediction, or brainstorming/mind mapping AT in postsecondary populations with LD.

Due to the limited number of studies examining the efficacy of AT to reduce academic deficits in LD postsecondary populations, the literature on AT abandonment was explored for its potential to indirectly inform on the usefulness of AT. The idea being, if AT abandonment is occurring at above average rates relative to its distribution, then this speaks to its efficacy, or lack thereof, in meeting the needs of a given portion of users. Notably, Todis (1996) reported that nearly one third of children and adults employing AT across educational settings abandoned the AT they had purchased or been provided.

In fact, the findings from a number of studies involving both children and adults with LD have provided a lengthy list of reasons related to why such individuals abandon the AT provided to them to help them manage their particular LD. The list of reasons includes the following: (a) the AT did not improve academic functioning, (b) the AT device was too difficult and expensive to repair, (c) the AT made the individual stand out in a group, (d) the AT required too much assistance from another person, (e) the AT was too difficult for the student to use, (f) the AT required a long or complicated series of commands, (g) the AT failed to function as intended, (h) the AT was not always reliable, and (i) the AT did not always address the curricular demands (Anderson-Inman, 1999; Chandler, Czerlinsky, & Wehman, 1993; Cunningham & Coombs, 1997; Raskind & Scott, 1993; Sherer, 1998; Todis, 1996).

Roessler and Kirk (1998) examined AT abandonment specifically within a postsecondary sample and deemed two factors to be associated with students’ abandonment of AT. First, the abandonment of AT was related to student perception that AT did not address their current academic needs. Second, due to limited AT training and supports, the postsecondary students expressed a general consensus of being unprepared to benefit from the technology. In a similar vein, Raskind and Higgins (1994) noted that if postsecondary students had to spend an inordinate amount of time learning to use technology they might abandon it and return to traditional nontechnological academic assistance in the name of short-term efficiency. More recently, Stodden, Roberts, Picklesimer, Jackson, and Chang (2006) surveyed disability support coordinators at 615 PSE campuses in the United States and found that while staff felt an acceptable level of technology and supports were provided to students with disabilities, AT task analysis/context evaluations were rated as the least available service. In fact, 58.5% stated that their institution did not provide such a service. The aforementioned makes clear the need for additional research pertaining to the interaction between the characteristics of AT users and their academic task demands to better understand the factors influencing acceptance or abandonment of AT.

The number of research studies available to review as a means to understanding the effectiveness of AT within a population of postsecondary students with LD is extremely limited at the present time. The number is so restricted in fact that, as far as can be discerned from attempting to find peer-reviewed articles, AT use to circumvent writing deficits has not been studied. It appears that the majority of research to date has focused on AT and its ability to circumvent reading disabilities. Regrettably, most of the research to date suffers from basic methodological flaws such as failure to consider LD subtypes and the resulting impact on sample composition and generalizability of results; small sample sizes resulting in low statistical power for analyses; and failure to use control samples. In addition, disability service provider differences in training and their availability for support have not been accounted for in any of the studies evaluating the efficacy of AT across institutions despite the fact that this could very well influence the training imparted to students with LD. Finally, perhaps the greatest impediment to advancing our understanding of the efficacy of AT in minimizing the effects of LD is the atheoretical approach taken by researchers.

Conclusion and Future Directions

Much remains to be done in this field of study. For instance, there is no published record of research assessing the efficacy of voice recognition, word prediction, and brainstorming/mind mapping software programs for postsecondary students with LD. Although the delivery of AT services on PSE campuses seems to favour the distributive location model (Abreu-Ellis & Ellis, 2006; Burgstahler, 1992), the proof for its effectiveness remains elusive as financial agendas and concerns regarding stigma appear to have driven this decision more so than evidence-based research. Research into this area with a focus on Canadian postsecondary institutions would be especially helpful as interprovince variations in levels of support exist. For instance, many of Ontario’s PSEs provide a level of service that exceeds what was given prior to the LOTF initiative, an initiative that has not been duplicated in other provinces. The examination of AT service delivery models is also dictated by the rise in Universal Instructional Design (UID) pedagogy within postsecondary institutions wherein AT is woven into the curriculum, assignments, as well as tests and exams.

Current definitions of LD conceptualize this disability as a deficit in a psychological process that leads to an academic impairment in the presence of other average abilities for thinking and reasoning (Learning Disabilities Association of Ontario [LDAO], 2001; Learning Disabilities Association of Canada [LDAC], 2002). Consequently, future research would benefit from examining how the AT used by postsecondary students with LD is able to tap into or replace the impaired psychological processes, which in turn influence the basic academic functions of reading and writing. Future research in AT is likely to advance the field if studies are designed which query not what academic skills AT is replacing but rather the psychological processes it is believed to be replacing (which in turn influence academic skills). This research approach as first used by Holmes and Silvestri (2009) frames questions in the following manner: If reading deficits are the result of core impairments in working memory, for example, then how well does AT circumvent working memory deficits and thus improve reading? Ultimately, we may come to learn that AT is a better substitute for some of these processes than others and thus able to improve the reading of only some students with LD.

Using the cognitive prosthesis metaphor to describe AT applications requires a change in the statistical analyses of AT efficacy research. With the use of a time series design (Smith, 2000; concurrent and differential approach), as opposed to traditional examination and comparison of test score means, research can project an individual’s performance on an academic task with and without the AT. This type of statistical analyses can predict the gap between performance with and without AT and forecast whether this gap will narrow perhaps to the point of insignificance. If such a trend were substantiated, then AT can be viewed within the cognitive partner metaphor, removable when the individual no longer needs it rather than as a permanent prosthesis. Whether this trend would occur, for which cognitive processes, and under what academic conditions is a beneficial line of future inquiry.

Currently, research investigating the efficacy of AT at the postsecondary level of education focuses on the examination of one specific type of AT to circumvent a student’s academic deficits. A possible future research agenda could take into consideration a broader notion of assistance, for example, studying combinations of AT or the pairing of AT with specific learning strategies to address individual student needs. The constellation of “high tech” (i.e., AT) and “low tech” (i.e., learning strategies) assistance is particularly advantageous to study in the current UID postsecondary environment, which fosters such combined approaches to teaching and learning.

The research on AT abandonment suggests there is a need for postsecondary institutions in Canada to conduct comprehensive and systematic evaluations of the AT needs of students with LD to measure the factors that interact with successful deployment of such technology. For instance, individual preferences, the range of tasks for which AT will be used, context in which AT will be used, and basic device qualities such as cost, portability, and aesthetics may all hold a role in the decision to routinely incorporate such tools into one’s life. Unfortunately, AT evaluations are at best endeavours based on trial and error, intuition, and completely dependent on the experience and training held by the evaluators, which varies considerably by province and territory in Canada. Furthermore, the existing, standardardized AT assessment protocols were not specifically designed for individuals with learning disabilities, but for individuals with a full range of disabilities (Raskind & Higgins, 1998; Sherer, 1998; Wissick & Gardner, 2008; Zabala et al., 2000) and so their utility must be considered as unproven for use with adults with LD. However, a review of technology assessment protocols by Wissick and Gardner (2008) concluded that three assessment models hold promise in addressing the needs of postsecondary students with LD: (a) the Functional Evaluation of Assistive Technology model (FEAT; Raskind & Bryant, 2002), (b) the Matching Person with Technology model (MPT; Sherer, 1998, 2005), and (c) the Student Environment Task Technology Framework (SETT) framework (Zabala, 1995, 2002). Future research is required to compare and contrast such technological assessment models when used within a postsecondary environment.

Psychologists when formulating recommendations around intervention are obliged by their national code of ethics and the standards issued by their licensing bodies to consider both the needs and characteristics of their clients as well as the theoretical and empirically supported efficacy of the interventions (e.g., Canadian Psychological Association, 2000; College of Psychologists of Ontario, 2009). Thus the state of knowledge with regard to the efficacy of AT to circumvent academic deficits as reviewed within this article has direct implications for the psychoeducational assessment reports written by psychological practitioners working with adult students. It is not at all uncommon for the recommendations contained within such reports to refer to AT as an accommodation for deficit skill development in the areas of reading and writing. Clearly the literature does not provide support for such a practice, but then again, neither does it clearly rule it out. Rather the strongest conclusion that can be generated at this point in time is that AT software may be of assistance to some students. It is advisable therefore for practitioners to couch recommendations for the use or purchase of AT with language of a nonprescriptive nature so that the recommended AT is not portrayed as a definite solution to the student’s academic deficits. Furthermore, because no research has been conducted comparing the efficacy of different brands of AT software, it is prudent for practitioners to recommend a class of technology, for example, “text-to-voice software,” rather than a particular brand name such as “Kurzweil™,” as possibly being of benefit to the student. Finally, because the factors associated with AT benefiting students are varied in nature and not fully known to date and because specialized evaluations of AT match to students is an unregulated and unstandardized practice, it is suggested that practitioners incorporate this notion into the recommendation section of their reports such that it is made apparent to the reader of the report that the usefulness of a type of AT software is likely best determined on a case by case and trial and error basis. In other words, students are encouraged to pursue the AT software and make their own determination as to whether or not it is truly of use to them given their particular academic deficits and the demands of their current program of study.

The use of AT by PSE students with LD appears to have occurred without research evidence validating or even testing the effectiveness of ATs in supporting the learning needs of this population. The collective body of research on the efficacy of AT to circumvent academic deficits at the postsecondary level of education is currently quite limited. Searches of peer-reviewed databases using a fairly exhaustive list of terms (e.g., AT, Adaptive Technology, Technology, computerized reading, screen readers, word prediction, speech synthesis, adult LD, postsecondary LD, college, university, Text Help™, Kurzweil™, Dragon Naturally Speaking™, etc.) drew only a handful of articles, all of which were considered within this review. The evidence available at this time is far from conclusive, as researchers have found that AT both hinders and helps adults with LD in their quest to overcome the academic deficits in the area of reading while insufficient work has been done to comment on the effect of AT on writing. This is likely the case as the extant research is so limited in quantity, conducted outside of any theoretical framework, and poorly controlled in terms of subject characteristics and delivery of AT. A promising avenue for future studies lies in the construction of methodologies based on a modern definition of LD, which identifies psychological processing deficits, such as working memory or phonological processing, as some of the causal mechanisms beneath disabilities in reading and writing (see Higgins & Raskind, 2005; Holmes & Silvestri, 2009). Until such shortcomings are addressed, it is unlikely that the assignment of AT to adults with LD will be based on much more than past practice and good will.