As explained earlier, ID imposes cognitive difficulties in reading, writing, spelling and the processing of numbers (American Psychiatric Association 2013b). Learners with ID typically have impairments in the following cognitive areas:

Language, communication and auditory reception.

Edyburn (2004) has posited that one of the primary reasons for learners to be transferred to a special school is because of the difficulty they have with reading. More specifically, these affected learners struggle to meet the requirements of the reading and viewing learning outcome. Learners with ID struggle to memorise and rehearse text that they have read. They also struggle to organise text or instinctively elaborate on it to assist them to learn (Alfassi et al. 2009). Learners who have reading challenges also generally have difficulty in phonemic awareness and analysis, word identification, reading fluency and understanding of the text (Elder-Hinshaw et al. 2006). This brings us now to the literacy specific needs and challenges of such learners.

Reading and writing constitute modalities through which language can be taught to young learners (Carstens 2013). Language can therefore be considered a tool for the conceptualisation and transfer of meaning and interpretation of texts (Silliman, Buttler & Wallach 2002; Wong, Graham, Hoskyn & Berman 2008). Reading according to Lessing and De Witt (2002) is a single aspect or learning outcome in literacy competence which can be described as the construction of meaning for which the learner must attain a necessary level of decoding proficiency. Children learn to read by progressing through a number of developmental processes. These are letter and word recognition, decoding, comprehension as well as how fluently the learner engages with the text (Long & Zimmerman 2009).

Word recognition refers to the ‘instant recall of words in which the reader resorts to no obvious mechanisms to recognize the word’ (Wong et al. 2008). When a learner is able to recognise words without hesitation, they have developed a state called automaticity that enables the reader’s brain to quickly and automatically process the words. In addition, learners who can instantly recall the words being read are capacitated with building mental representations of the message of the text that they can tap into when required to demonstrate their comprehension of a story (Allor et al. 2010).

Reading comprehension is the ability of the child to understand the text being read (Lea & Street 2006). Learners with reading difficulties generally experience poor comprehension because of their lack of being able to read and monitor their understanding of the information. The ability to decode words, poor vocabulary access and fluency in a language are contributory factors to a learner’s comprehension (Wong et al. 2008).

Reading fluency, being the last of the processes, relates to the speed and accuracy in the execution of the reading task (Chard Vaughn & Tyler 2002). The pace of a learner’s reading affects the way the learner retains the information and develops meaning from the text (Wong et al. 2008). Therefore, learners who read at a very slow pace struggle to retain the information being read, and consequently develop inadequate meaning from the text.

For a learner with ID, interventions that entail the teaching of language processes described above have been found effective in improving their reading abilities (Al Otaiba & Hosp 2004; UNESCO 2006). Given the literacy challenges, for example word identification, reading fluency and comprehension of the text, that the intellectually disabled learner has to contend with, what strategies can educators employ to ameliorate the literacy of such learners?

According to Edyburn (2004), when learners struggle to read, educators tend to resort to other methods of instruction. The problem is that a new method of instruction may not necessarily yield positive results. This is particularly true if the very problem of reading is part of the inherent nature of their disability. It is our view that, if one holds the position that all learners can learn, albeit at different levels, then any support mechanism or strategy would be worthwhile (DoE 2001; Gindis 2003). Among well-documented strategies is the use of assistive technology which is defined as ‘any item, piece of equipment or product system, […] used to increase, maintain, or improve functional capabilities of individuals with disabilities’ (Hobbs et al. 2009:153). The Foundation for Assistive Technology in the United Kingdom extends this definition to the ability to enhance ‘independence for disabled’. Brodwin, Star and Cardoso (2004:29) indicate that assistive technology does not only involve computers with all their components but involves an ‘integral process of assisting individuals with disabilities, […] to maximize their human potential’. Furthermore, computer assistive technology has due consideration for the learner and his or her individual traits, as well as his or her abilities and challenges. There are hosts of computer software on the market that can meet the specific needs of different groups of persons with a disability. In the section that follows, we highlight some of the technologies that have been used to aid learners with intellectual disabilities.

Thompson (2005) refers to text-to-speech technology as a type of multimedia program that has the functionality of converting computer text to digitised speech. Zhao (2007:35) in turn indicates that speech technology refers to ‘technology that enables machines to receive and accept human oral language as input and respond with human or human-like oral language as output’. An important element of this technology is that it enables the learner to access software applications or content with immediacy in the speech feedback that could allow the learner to correct reading mistakes. Forgrave (2002) and Zhao (2007) are of the opinion that speech technology minimises the decoding problems that disabled learners sometimes have, which allows for better comprehension. Furthermore, it provides learners with repetitive visual and auditory cues that can help them to comprehend the text.

E-text can be defined as the ‘text that has been altered to increase access and provide support to learners’ (Edwards 2008:36). Supportive electronic text aids learners with disabilities in dealing much better with text with the use and support of computer technology. One of the advantages of the computer software is the ability to change the way text is viewed and read, by modifying the font size and colour. The text can also be read aloud. Further to this, multiple images can be shown at any given time (Anderson-Inman & Horney 2007). These features are in contrast to printed text. The printed material in general does not afford the reader the opportunity to customise the text being read to them. Interestingly enough in instances of electronic text and printed text, the educator has to play a supportive role as well.

E-books serve to replicate the printed paper-based storybooks into a digital format. In contrast to print paper-based books, e-books have additional multimedia effects to support the learner’s understanding of the text (Rhodes & Milby 2007). E-books have a number of intrinsic elements such as sound, animation and interactive activities. These elements can scaffold the learning of the learner, allowing him or her to eventually master the given task. In the process, the learner can be exposed to chunks of the reading task, for example to read one paragraph a few times then asking the learner to retell that part.

The interactive nature of e-books makes them particularly very attractive for young learners, and they tend to repeat activities which increase learning (Picton 2014). Learners with special educational needs, such as struggling readers, can therefore benefit from the additional text features of electronic books (Larson 2010). This particular feature could prove to be helpful for the participants of this study. The learners’ intellectual disabilities as has been indicated earlier range from mild to severe. This implies that the learners’ reading ability levels also vary. Multiple opportunities to expose the learners to text, which they have experienced on an auditory and a visual level, could aid their reading ability. ‘E-books and other text-to-speech readers boost students’ self-esteem while providing access to texts that were previously out of reach’ (Rhodes & Milby 2007:256).

Furthermore, children’s books that have been recreated into an electronic format allow the learner to track print and view a visual representation of the story. Electronic books help the learner to build their vocabulary, aid the understanding of the text while at the same time showing them how to read fluently (Horney & Anderson-Inman 1999; Rhodes & Milby 2007). However, some limitations do exist in the use of talking books. For example, e-books have limited use in the classroom during the course of the delivery of the literacy curriculum. Despite existing evidence from Chera (2002), some researchers are of the opinion that the true value of talking books having real educational potential has yet to be realised (Fox 2002; Littleton, Wood & Chera 2006). Research conducted by Chera (2002) has shown that talking books can promote phonological awareness in children during their initial reading experience of learning to read. Learners that have well-developed phonological awareness recognise on an auditory level that words can rhyme, start or end on the same letter and that letters can be manipulated to form new words. This unlocks future developmental skills that allow the learner to reflect and manipulate letters to create new words (Marthinussen 2011). Phonological awareness is regarded as a significant antecedent skill to the successful acquisition of reading (Adams 1990b; Blachman 2000; Goswami & Bryant 1990; Littleton, Wood & Chera 2006). Accordingly, the reader has to learn to master phonological skills that will enable him or her to break up the speech or the spoken word into phonological segments. It is clear from these studies that talking books have the potential to support reading development.

Teaching children with intellectual impairments entails the use of appropriately identified technological aids. Text-to-speech technology is one such technologically intensive teaching aids in the context of special needs teaching and learning. This does not however mean that the decision to employ technology is taken blindly. We consider and accept the caution made by Roulstone (2016) that the use of technology, especially in special needs education, should be couched in the context of the learners. In the case of this study, the learners and the teachers are bilingual in English and Afrikaans. The use of text-to-speech technology is therefore suitable and easy to achieve because both English and Afrikaans are official languages in South Africa and are heavily used in many educational platforms. This consideration was important to negate the possibility of blindly adopting technological tools for the sake of technology, or to provide false hope of the potential of the tool to the learners (Breen 2015; Marchal-Crespo & Reinkensmeyer 2009).

This action research took place at a special school within the Cape Metropole area that caters for learners with ID. The primary convenient sampling group comprised 35 learners across the five intermediate phase classes of the school where one of the researchers worked. To ensure that none of the learners was put at an advantageous or disadvantageous position in relation to one another, the following measures were undertaken: (1) all learners were exposed to the same reading activity, whether dependent (active educator support) or independent (less active educator support) and (2) no undue attention was granted to anyone beyond the level of support that would have been granted to all in the class during class activities. In this sense, the use of a control group was not an option so as to keep with ethical promises that all learners would be equally exposed to potential benefits of the intervention.

In addition to being in the intermediate phase (Grades 4–6), the learners met the following criteria:

Prior to their participation in the study, the learners had been diagnosed as being intellectually disabled (the diagnosis of ID was carried out prior to placement at the special school) and were chronologically between 10 and 14 years old with an average mental age of between 4 and 7 years old.

The secondary sample group comprised the four educators of the learners. The reasoning behind the selection of this group was based on their daily direct involvement with the learners. The function of the educators was to assist the researchers to perform the pre- and post-tests as well as to engage the learners during the intervention sessions. The educators’ role in the study was therefore to provide qualitative accounts of learners’ progress or lack thereof.

The school selected for this study had the following contextual characteristics:

well-resourced school (computer equipped with a computer lab and Internet),

Semi-structured interviews were conducted with the four educators helping with the tests’ administration and the intervention. The main purpose of the interviews was for the educators to share observations they had made about learners during the two processes. The interview questions were as follows:

What in your view is the role of technology, especially computers, in the learning and teaching of mentally challenged learners?

The study received ethical clearance through the CPUT Faculty of Education Ethics Committee (Approval Certificate EFEC 2-7/2009). Permission to conduct the study at Western Cape Public School was obtained from the Western Cape Education Department Research Directorate.

Two clearly defined groups emerged when performance was analysed. The first learner group are those who managed to read between 4 and 35 words, compared with the second group that had a very low to zero (0) reading pre-test score. From the outset it appeared that the strides made by learner group 1 (LG1) (high pre-test) was much higher than that of learner group 2 (LG2). That is, the learners who were able to identify more words during the pre-test were able to replicate and improve on that score during the post-test. In addition, time spent on reading (speed) by the LG1 was between 27 s and total given time of 60 s. The LG2 spent a full 60 s on the reading tasks. Beyond the question of how many words were read and how many errors were made, the next level of analysis focuses specifically on: (1) learners’ reading behaviours and (2) whether they made or did not make errors.

The behaviours and errors during reading were categorised along language processes and milestones expected of learners as they learn how to read. These are ‘letter and word recognition’, ‘decoding’, ‘comprehension’ as well as how fluently the learner engages with the text (Long & Zimmerman 2009:4). The categorisation was based on a constant comparison of all behaviours and errors. A feel or look alike criteria suggested by Maykut and Morehouse (1994) was used during the constant comparison to determine the categories. This process can be likened to the ‘first reading of, or first order imposition of meaning, on the data’ conducted in order to arrive at specific categories of unfolding descriptions of learner performance (Mosito 2005). The description of errors and reading behaviours follows.

In line with the above descriptions, Table 1 depicts the performance of LG1.

It is clear from Table 1 that the most dominant feature of LG1 reading behaviour was problems with word and letter recognition. On the positive side, there was a notable decrease in word and letter recognition problems during the post-test. In addition, during the post-test some of the learners (3) indicated comprehension of the story by adding correctly to the context words(s) which were not in the story they were asked to retell.

The positive behaviours noted in Table 1 are corroborated by educators’ observations as they responded to questions on (1) general achievements and (2) specific reading behaviours with regard to reading processes such as word recognition, comprehension and fluency. On the first question, there was consensus among the four educators that computer technology seemed to facilitate independence and discipline. One of the participating educators commented,

‘I pick up a tremendous enthusiasm in the learners … they insist they want to read every day … they longed for what they have learnt … means a lot for them … I must just make it smaller … they must get the exercise everyday… when we go to the computer lab … they want to go straight to the stories … good thing [Text translated from Afrikaans].’ (Educator 4, female, LG1)

In Table 2, findings from the reading performance of the remaining 26 learners are presented.

Similarly to the LG1 in Table 1, LG2 also demonstrated high instances of word and letter recognition difficulties. The incidence of these difficulties improved slightly during the post-test. Five of the learners demonstrated comprehension of the story through their ability to add a correct word to the initial story. One of the observations made by the researcher-educator was that the errors that were made by LG2 were in many ways far removed from the real word or were just nonsensical. These types of errors were based on the fact that the learner could not decode the requested word. This observation explains the slight decrease in errors noted between the pre- and post-tests.

As the study was not about a performance of a particular group against the other (LG1 vs LG2), in Table 3 all the behaviours per category from both groups was added in order to arrive at the overall picture of all the learners. These totals inform the discussion that follows.

It is clear from the above overall depiction of learner performance between the pre- and post-tests that there was a positive upward movement in learners’ comprehension of the story. The positive trend is indicated by a high level of meaningful insertions of words which did not feature in the actual story words and the learners’ ability to retell the stories. Educators’ observations of learner performance were in line with the trend. For example, Educator 3 disclosed

‘… Word recognition, … sight reading … and now I see with some of them one could move beyond that; visual discrimination of words, … showing me word for word, … using that method, … found that it works well for me.’ (Educator 3, male, LG1 & 2)

Another important observation made by the educators was that the text-to-speech artefact had been the most exciting part of the interventions. Educator 1 explained that stories told and seen through the computer seemed to generate an interest level she had not seen in learners before. In her own words, ‘a child views a story on the computer, hearing this voice coming from the computer … makes learning interesting and fun’ (Educator 1, female, LG1 & 2).

With the picture portrayed by the findings in Table 3, we are now at an informed position to address more specifically the question of the nature of change in reading abilities learners with ID in this study have undergone following teaching that was mediated with text-to-speech books.

The biggest reading difficulty experienced by learners in this study pertained to word recognition. Allor et al. (2010:4) state that ‘good readers effortlessly recognize words and build mental representations of the message of the text’. This implies that learners who struggle with word recognition will be robbed of the tools for comprehending and storing messages for later recall. This finding leads to a conclusion that word recognition is a major challenge for intellectually disabled learners. While there appears to have been improvement among one group of learners in this instance, the duration of the intervention might not have allowed for high levels of achievement across all participating learners.

Where reading comprehension is concerned, the performance of learners in this study confirms what previous research has revealed. Learners with word recognition difficulties have lower grasp or comprehension of the story which is often exacerbated by poor decoding required for word recognition (Wong et al. 2008).

A number of correct words and sentences read and time spent on reading are indicative of reading fluency which is defined as the speed and accuracy in the execution of the reading task (Chard, Vaughn & Tyler, 2002). Another finding in this study is the extent to which the participating learners were challenged in this process and improved or did not between the pre- and post-tests. The analysis resulted in two distinct groups of learners in this regard (LG1 and LG2). The former group’s performance marked them as the stronger lot as they utilised far less of the provided reading time and some managed to read all the 35 words. Most of the participants in LG2 utilised all the given time of 60 seconds but failed to read a single word correctly.

It is on the question of the emerging reading processes that one limitation of the study reveals itself. Initial in-depth analysis of data should have occurred after the pre-test in order to isolate those areas that the intervention could have addressed. This limitation is a crucial finding by itself which future research could address. The second limitation relates to the absence of a control group. It is possible that what we have interpreted as reading gains could have resulted from the use of any teaching artefact besides the text-to-speech stories. Performance of a control group which was not exposed to the intervention could have led to clearly established conclusions. Lastly, that the analysis isolated two groups and this was not followed up, for example in terms of probing from the educators concerned if they did anything differently which in their opinion could have resulted in higher gains for LG1, was a lost opportunity for revealing the specific variables that could have added to this marked difference between the two groups.