Information and communication technology reading interventions : A scoping review

The majority of South African learners are not developing the reading skills expected for each grade when compared to their international peers (Department of Basic Education 2014; Mullis et al. 2017). This is a multifaceted problem linked to a complex interplay of educational, political, social and economic factors described by authors such as Spaull (2013). Factors related to learners’ underperformance include resource constraints, inadequate teacher training, poor instructional practices, low parental literacy levels, learning in a second or additional language, and high rates of absenteeism (Howie et al. 2017). Multifaceted interventions that focus on a range of aspects such as infrastructure, teacher training and classroom interventions are necessary to address the situation and bring about a more positive outlook.


Introduction
have demonstrated effectiveness, as well as to expand the evidence base through ongoing intervention studies.
Reading intervention research shows that targeting phonological awareness, phonics, fluency, vocabulary and reading comprehension in an explicit, intensive and systematic manner improves reading skills (Galuschka et al. 2014;Gibson & Musti-Rao 2017;National Reading Panel 2000;Suggate 2016). Despite this knowledge, serving the large population of children requiring reading intervention is a challenge, especially in resource-constrained contexts. One approach to providing reading support is through the use of information and communication technology (ICT) programmes. There is a wide range of technology-based tools available to help children develop their reading, spelling and language abilities. These vary in terms of the ages targeted, their specific focus, platforms used, accessibility and cost. Although not all studies investigating the impact of ICT-based approaches to reading have reported positive outcomes (Campuzano et al. 2009), reviews and meta-analyses indicate that many ICT programmes produce gains in phonological awareness, phonics, word reading, fluency, vocabulary and reading comprehension of schoolaged children (Cheung & Slavin 2011Jamshidifarsani et al. 2019;Moran et al. 2008). Information and communication technology-based reading intervention holds potential for improving the reading skills of children by harnessing their motivation to learn through feelings of autonomy (making choices), competence (achieving goals), and relatedness (sharing experiences with another individual), providing immediate feedback and having the capacity to be intensive, individualised, and at the appropriate level of difficulty, and enabling independent use or the presence of nonprofessionals (McTigue & Uppstad 2018).
In this scoping review we set out to describe ICT interventions for reading and their outcomes as described in the literature. A particular aim of the study was to consider interventions developed for, or investigated in the majority world (developing) contexts which might offer solutions to the challenges faced in South Africa.
The objectives of the study were to: (1) provide an overview of ICT-based reading interventions described in the literature over the last decade (2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019) and (2) consider the subset of ICT reading interventions conducted in the majority world and their impact on learners' reading skills and challenges faced, which could lead to recommendations for research conducted in similar contexts, such as South Africa.

Methodology
Scoping reviews are used to map the main sources and types of evidence available, and are particularly useful when an area is complex or has not been reviewed comprehensively before. Arksey and O'Malley's (2005) framework has five steps: (1) identifying the research question or aim. This review set out to describe ICT reading interventions for primary school learners undertaken in the last decade.
In particular we wanted to know what work has been undertaken in the majority world so that we could investigate programmes reported to be effective in this context and build on them further. (2) Identifying and (3) selecting relevant studies. A search strategy, criteria for eligibility and study selection were devised, and are described in the following sections. (4) Data are then charted, collated and (5) reported in the results section of the article. The search took place between June 2018 and June 2019, undertaken mainly by the first author with the other two authors in a checking and support role. To ensure a valid and reliable process, measures were put in place such as team briefings on a regular basis to discuss any uncertainties regarding the process and findings to date.

Search strategy
First, a pilot phase was initiated in which one database was searched using a set of core terms. Titles, keywords and index terms taken from this initial set of papers were then used to develop the list of search terms further. Second, following the pilot phase, researchers then used the complete search term list with the full set of electronic databases. Keywords were entered into the electronic databases of PsycArticles, PsychINFO, ERIC, Computers and Applied Sciences Complete, Academic Search Premier and CINAHL.
The keywords were: information and communication technology; computer-assisted; computer-based; laptop; smartphone; iPhone; tablet; iPad; application; programme; software; reading intervention; reading instruction; reading therapy; reading remedial; primary school; elementary school; middle school; junior school; children and learners.

Eligibility criteria
Studies were included in the review based on the following inclusion criteria: • Published in a peer-reviewed journal between 2009 and 2019. • Interventions described needed to be delivered by ICT, and aimed at improving reading or reading-related skills (one or more of the components of phonological awareness, letter-sound knowledge or phonics, word reading, fluency, vocabulary and reading comprehension). • Learners in Grades 1 to 7 were the participants. • Experimental or quasi-experimental designs were used, that is, the included studies all considered the effect of an intervention on particular outcomes; control groups were used, although in the case of a quasi-experimental design assignment into the groups was not random.
Due to time and resource constraints we were only able to access and review papers in English, and grey literature (e.g. postgraduate student projects, government reports) was not searched. Meta-analyses, reviews and editorial or discussion pieces were excluded. We wanted to access original research papers that might have contributed to a

Overview of ICT-based reading interventions described in the literature
In the 49 papers included in this part of the study, the most commonly used design was an experimental pre-post design with random assignment to groups at the level of schools, classes or learners ( Of the studies that reported the language characteristics of the participants, 29 investigated monolingual children, 9 involved bilingual children and 4 worked with bi-and monolingual children. Many studies did not report on the languages of the participants. A total of 46 different ICT reading interventions were described. Some studies evaluated one programme while others used two or more, comparing outcomes between groups. Programmes used in more than two studies included GraphoGame/GraphoLearn (henceforth GraphoGame) (used in 10 papers), ABRACADABRA (6 papers), Reading RACES (3 papers) and Chassymo (3 papers). These 'big four' programmes thus dominated the literature for the time period investigated. GraphoGame targets multiple levels of reading (phonics and letter-sound knowledge, phonological awareness and word reading). It is a theoretically informed intervention that has been well researched over many years and adapted for use in a variety of different languages. The programme is available to all school-aged children in Finland, and in many other countries around the world in adapted forms (see Ojanen et al. 2015 for further information).
ABRACADABRA similarly targets a range of skills including phonics and letter-sound knowledge, word reading, reading and listening comprehension, reading fluency and metacognition in reading and writing. It was developed in Canada and has been extensively used there as well as in Australia. It is based on the recommendations of the National Reading Panel (2000) Table 1 shows a mapping of the 49 papers to give an overview of the designs used, sample sizes and participants, intervention and outcomes. Further detail for each of the 49 papers is provided in the appendices (Tables 1-A1-3-A1).
Results from the first part of the study indicated a substantial number of ICT interventions for reading that have been researched and published in peer-reviewed journals over the past decade. Most of the programmes demonstrated positive   effects on children's literacy and language, despite variation in the nature and duration of the programmes. A small set of programmes (GraphoGame, ABRACADABRA, Reading RACES, and Chassymo) appeared in multiple studies and seem to have been most rigorously investigated in the 10-year period to demonstrate their effectiveness in different contexts. Given our specific context and challenges in South Africa, the second part of the review focused on a subset of studies from the main scoping review, ICT interventions undertaken in majority world countries, which we considered might be especially applicable to children in South Africa.

ICT reading interventions conducted in the majority world and their impact on learners' reading skills
In the main data set we found five studies conducted in the majority world: two in Kenya, and one in each of Zambia, India and Tanzania, constituting 10.2% of the total number of papers found in the review. These five studies used two interventions from the 'big four' group introduced in the previous section: ABRACADABRA and GraphoGame. All studies conducted in the majority world demonstrated improvements in learners' reading skills as a result of the ICT intervention. Two of the studies found positive intervention effects on all outcome measures and three noted positive intervention effects on at least half of their outcome measures. Table 2 provides an overview of these studies. Lysenko et al. (2019) conducted a study in Kenya that examined the effect of English ABRACADABRA intervention (targeting phonological awareness, phonics, word reading, fluency, vocabulary, listening comprehension, reading comprehension and writing) and READS intervention (online stories and books available in English and Kiswahili to improve reading fluency and comprehension) on English literacy skills. A quasi-experimental design was used with a large sample (n = 1672) of Grade 1-3 children learning English as a second language in mainstream schools. Schools were non-randomly assigned to ICT intervention or control conditions. Trained teachers facilitated intervention for a total of 2 h per week for a total of 16 weeks. Standardised assessments of English oral language and reading skills as well as participants' national examination results (in English, Social Studies, Mathematics and Science) were used as outcome measures. The results showed that intervention participants significantly outperformed control participants on all measures. Another study using ABRACADABRA was undertaken by Abrami et al. (2016) in Kenya to investigate the impact of English ABRACADABRA intervention on mainstream Grade 2 children (n = 354) learning English as a second language. An experimental pre-test-post-test design with random assignment of classes to conditions (intervention versus no intervention or control) was used.
Intervention participants made significant gains in reading comprehension and listening comprehension compared to the control group and participants in the intervention group significantly outperformed children in the control group in the national examinations.
Ngorosho (2018) conducted a study in Tanzania that investigated the impact of Kiswahili GraphoGame (targeting phonological awareness, phonics and word reading) on Kiswahili literacy skills. An alternating treatment design with random assignment of schools to groups (ICT intervention versus non-ICT classroom intervention versus no intervention or control) was used. Participants were Kiswahili home language Grade 1 learners (n = 49) with poor reading skills. Participants accessed GraphoGame via smartphones and worked independently (without adults being involved in instruction) for three sessions per day, 10 min per session, 5 days a week (a total of 2-4 h of intervention). Non-standardised outcome measures were used. The findings indicated significant improvements for both the ICT and non-ICT classroom intervention, although the ICT intervention led to the greatest improvement.
Jere-Folotiya et al. (2014) conducted a study in Zambia to determine the effect of ciNyanja GraphoGame on mainstream Grade 1 (n = 573) children's ciNyanja literacy skills. As for the Tanzanian study, learners accessed intervention independently on smartphones 3-5 days per week (for six sessions which were 7-9 min long per day) for a total of 1 h and 34 min of intervention. Standardised measures of orthographic awareness and spelling acted as outcome measures. The results showed that the intervention improved the spelling (intervention participants significantly outperformed control participants) but not the orthographic awareness of participants. The learners who were exposed to intervention directly (played GraphoGame) and indirectly (teacher played GraphoGame) produced significant improvements in spelling compared to control learners.
Patel et al. (2018) conducted a study in India using an experimental pre-test-post-test design with random Mainstream learners assignment to conditions (ICT reading intervention versus an ICT Mathematics intervention control). Grade 3 learners (n = 30) with reading difficulties who did not have English as a home language but were attending a school where English was the medium of instruction participated. Intervention participants used English GraphoGame (to improve phonological awareness, phonics and word reading) independently on tablets for 8 weeks (20-30 min per session and six sessions per week). Non-standardised (informal ingame assessments) and standardised literacy assessments were used as outcome measures. Significant improvements in favour of the intervention group were found for all GraphoGame in-game measures but there was no difference between the improvements of the intervention and control groups on the standardised measures. These interventions are summarised in Table 3.
The five studies conducted in the majority world (Abrami et al. 2016;Jere-Folotiya et al. 2014;Lysenko et al. 2019;Ngorosho 2018;Patel et al. 2018) reported a range of study-related challenges: learners had limited exposure to technology before beginning the intervention, high rates of learner absenteeism, learners arriving at school late, finding time for supplemental ICT intervention in a curriculum-determined timetable, lack of linguistically and culturally appropriate assessment measures (and lack of standardisation of assessments on the study population), technological issues, and venue constraints (no quiet venues at schools where intervention and assessments could be conducted). They reported the following contextual challenges: poor infrastructure, shortage of reading and teaching materials, poor working conditions for teachers, teachers inadequately trained for literacy instruction, poor teaching methods (rote learning), learners not being exposed to the language of learning and teaching in the home environment, parents having low literacy levels or being illiterate, and lack of learner exposure to literacy activities in the home environment.

Discussion
Research conducted in the last decade suggests that ICT approaches to reading intervention can lead to improvements in learners' reading skills, and thus offers potential for providing support to large numbers of children who require it -especially in places such as South Africa which has an acknowledged literacy crisis (Spaull 2013). There were only four studies in the scoping review that were conducted in Africa and none of these was conducted in South Africa. However, the small subset of papers from the majority world tentatively suggest that ICT-based approaches to reading intervention may be helpful for improving the reading skills of children in settings with a similar socio-economic status and achievement profile to ours. The review highlights a gap in research that should be addressed due to the potential positive impact ICT reading interventions could have in such contexts.
Speech and language therapy is an evidence-based profession so professionals recommending or facilitating ICT literacy interventions will need to know which interventions have been shown to be effective, and the extent of the evidence. Some of the interventions in this review have been researched in many studies including other countries in Africa. We described the 'big four' ICT interventions emerging from the review which were GraphoGame, ABRACADABRA, Reading RACES and Chassymo.
They are starting to emerge as programmes that could be considered for at-scale intervention in South Africa. In particular GraphoGame and ABRACADABRA have been used in many different contexts, including majority world countries. GraphoGame has been successfully adapted into http://www.rw.org.za Open Access many languages and this adaptation would be particularly important for its use in South Africa too. Here it would be important for the programme to be adapted into learners' home languages, as well as second or additional languages because in Grade 4 there is a shift from mother tongue instruction to English or Afrikaans instruction in schools.
Care would need to be taken to ensure cultural and linguistic appropriateness, but the effective adaptation and use in Zambia and Tanzania suggests that adaptation would be feasible. Home language interventions could be used to scaffold and support the development of English or Afrikaans language and literacy skills. The review found only a few studies with bilingual participants, which does not reflect the high international rates of childhood bilingualism (Paradis, Genesee & Crago 2011). There is also a great need for interventions to be developed specifically for bilingual learners and for their efficacy to be determined.
Although the potential of interventions such as GraphoGame for South Africa is clear, it is not to say that this intervention works better than other interventions. Rather it has been researched more and the process of evidence generation is more advanced than for some other programmes. The relative cost effectiveness and efficiency (ease of implementation) of each approach would need to be examined further so that the approaches with the strongest evidence and feasibility for implementation are selected at schools. Evidence-based practice refers to interventions and their application in particular contexts with particular individuals. Thus, interventions that are designed and tested in a particular setting with a group of children with particular characteristics may not be appropriate or effective for another setting or group of children. The use of established interventions would also not preclude the urgent need to develop and trial our own local interventions that may ultimately prove to be as effective.
South Africa has one of the highest mobile phone penetrations in the world (Ojo 2018) and therefore the use of smartphonebased interventions for improving both health and education is relevant. The two GraphoGame studies undertaken in Africa involved children using smartphones to access the intervention, and this approach may be worth considering further. During the COVID-19 pandemic when South Africa's schools were closed many children had no access to learning materials or any sort of educational support.
Smartphone-based apps such as GraphoGame could enable children to develop their reading skills anywhere and anytime, especially if the relevant apps were freely available and access to the data was zero-rated. Although findings related to intervention intensity and facilitation did not reveal clear patterns related to their relative effectiveness, studies with four to five sessions per week produced more widespread improvements than those with one to three sessions per week.
The studies conducted in the majority world reported a number of challenges associated with intervention delivery. Information and communication technology-based reading intervention studies conducted in similar contexts should provide training to learners on how to use technological devices and provide opportunities for learners to become familiar with the devices before intervention commences. Smartphones are likely to be familiar to many children, but they may not be able to have access to their own device and may not be able to bring it to school. Innovative solutions to venue constraints such as having multiple intervention slots where only a few learners attend an intervention at each time, rearranging furniture and using classroom dividers, and dividing learners into groups may need to be considered in schools -highlighting a need for addressing basic infrastructure, which remains problematic in many South African schools. There is an urgent need to develop culturally and linguistically appropriate reading, teaching and assessment materials together with training opportunities related to how resources can be used. Workshops for teachers that focus on effective literacy instruction, and for parents who have low literacy levels related to how they can support their children's literacy learning, will also be valuable and should be seen as part of a broader solution to addressing challenges that an ICT intervention alone will not be able to address.
The scoping review may present a biased impression of the effectiveness of ICT interventions as studies demonstrating no effect are less likely to be published (Djulbegovic & Guyatt 2017). Our review was also limited in that we only focused on a 10-year period and did not consider studies that were published in languages other than English and that used non-experimental methods. A systematic review or metaanalysis that seeks to address more specific questions about ICT reading interventions would be helpful, along with further studies that trial and evaluate ICT interventions in the majority world where they are most needed.

Conclusion
The scoping review described the characteristics of 49 ICT-based reading intervention studies for primary school children, published in the past decade. Findings indicate a promising range of different ICT-based interventions, most of which demonstrate positive outcomes although wideranging outcomes measures and research designs have been used. Only a small proportion of the studies were undertaken in the majority world. There is a great need for further work in this context and in particular in South Africa where reading outcomes are poor. It is clear that ICT-based approaches to reading intervention can lead to improvements in learners' reading skills, but further research is needed to determine if any of these interventions could be relevant for South African learners or to guide the development of innovative ICTbased interventions responsive to the needs of South African children and educators.
http://www.rw.org.za Open Access Authors' contributions J.D., M.P. and J.l.R. contributed to the design and implementation of the research, to the analysis of the results and to the writing of the article.

Ethical considerations
This article followed all ethical standards for carrying out research.  (6)

Funding information
Informal measures of word reading accuracy and fluency. All children's reading accuracy and fluency increased (the number of correct words read per minute) as a result of the intervention.
Standardised measures of reading and decoding accuracy, reading and decoding fluency, and spelling of words.
Phase 1: All groups improved significantly on all outcome measures. There was no difference among the performance of the three different intervention groups. Phase 2: All groups improved significantly on all outcome measures. There was no difference among the performance of the three different intervention groups except that decoding accuracy showed greater improvement for the phoneme-level intervention group compared to the rime-level intervention group. Barber et al. (2018) Informal measures of reading fluency and comprehension. Children improved in reading fluency and comprehension as a result of the intervention.
Messer and Nash (2018) Standardised measures of phonological awareness, phonological short-term memory, executive loaded working memory, naming speed, decoding and spelling.
The experimental group had significantly higher scores than the control group on measures of decoding, phonological awareness, naming speed, phonological short-term memory and executive loaded working memory.
Ngorosho (2018) Informal lexical decision tasks for letters, syllables and words and informal single word spelling test. Informal measures of letter knowledge and word reading.
Significant improvement was found for both GraphoGame and classroom instruction interventions. GraphoGame intervention led to greater improvement than classroom instruction intervention.
Patel et al.
Informal GraphoLearn in-game assessments: grapheme-phoneme knowledge, rime unit recognition, whole word reading. Standardised measures of word reading, sight word reading efficiency, pseudoword reading efficiency and spelling.
Significant differences in favour of the intervention group for all GraphoLearn in-game measures. No significant differences between groups on standardised measures. Both groups showed improvement on all measures from pre-test to post-test.
Solheim et al.
Standardised measures of word reading, sentence reading and spelling.
Both treatment groups had statistically significantly higher word reading, sentence reading and spelling skills than the control group at post-test follow-up 1 and 2 years later.
Baker et al.
Standardised measures of Spanish pseudoword and word reading, standardised measures of English pseudoword reading and English reading accuracy and fluency.
No significant difference in the gains on Spanish pseudoword and real word reading accuracy and fluency between the intervention and control conditions. No significant difference in gains on English reading.
Bennett et al.
Non-standardised measures of reading fluency and comprehension. Reading fluency and comprehension improved for practised and novel passages during intervention and on the 2 week and 1 month follow-ups.
Horne (2017) Standardised test of reading comprehension, reading rate and reading accuracy.
Significant effects for reading comprehension and accuracy but not for rate. The intervention group in school 2 (two sessions per week) demonstrated significantly larger gains in reading accuracy and comprehension than the group in school 1 (one session per week).
Standardised assessments of word reading, decoding, phonological skills, decoding fluency, receptive vocabulary and non-verbal reasoning. Informal measures of listening comprehension, reading comprehension, comprehension monitoring, and working memory.
Grapho-syllabic training led to improved performance in written word recognition (as well as phonological awareness and decoding). Comprehension training improved listening and reading comprehension (and vocabulary and comprehension monitoring).

Madden and
Slavin (2017) Standardised assessments of word reading, fluency and comprehension.
Reading skills improved significantly in intervention group compared to control groups.
Mak et al.
Standardised measures of word reading, vocabulary reading, listening comprehension, phoneme segmentation fluency, and phoneme blending fluency.
Learners in the intervention scored significantly higher than control learners on vocabulary, reading, phoneme segmentation and blending fluency. Both groups performed similarly for word reading, and listening comprehension.
Moser et al.
Standardised measures of vocabulary, reading comprehension, rate, accuracy, and spelling.
No intervention effects were found; intervention and control participants showed no significant differences in any areas.
Ozbek and Girli (2017) Informal measures of reading fluency. Intervention improved the reading fluency of the learners. Van de Ven et al. (2017) Standardised tests of pseudoword reading, word and text reading, receptive vocabulary.
Intervention effects were found for pseudoword reading and reading fluency.
Norm-referenced tests of reading, reading vocabulary, reading comprehension, listening comprehension and end-of-year examination results in English and other subjects taught including Social Studies, Mathematics and Science.
Intervention group made significant gains in reading and listening comprehension compared to the control group. No significant difference between the groups on other measures. Intervention group outperformed children in the control group in the exams (small effect size).
El Zein et al.
Informal measure of reading comprehension. Teacher-directed instruction was more effective than iPad instruction for increasing the accuracy of responses.
Significant improvement in reading and reading comprehension between pre-test and post-test measures. At a 1-year follow-up, the children in the intervention group had developed reading skills at a rate equivalent to typically developing children.
Intervention group improved significantly more than children in the control group on measures of word spelling, non-word spelling, and word fluency. Significant effects were not detected for non-word and real word reading but there were non-significant differences favouring learners in the treatment group.
Musti-Rao et al. (2015) Informal measures of sight word reading; standardised measures of reading fluency.
Improvements in sight word reading; gains not achieved for oral reading fluency.

Shannon et al. (2015)
A norm-and criterion-referenced assessment of word knowledge, analysing literary text, understanding author's craft, comprehension strategies and constructing meaning, and analysing argument and evaluating text.
The intervention group showed significantly greater gains on the assessment than the control group. A moderate effect size was found. The programme had a positive impact on learner's reading achievement.
Tyler et al.
Standardised measures of reading accuracy, fluency and word recognition.
Significant improvements in favour of the intervention group for reading accuracy but not for reading fluency or word recognition.
Jere-Folotiya et al. (2014) Standardised tests of orthographic awareness and spelling. Intervention improved the spelling of participants (intervention learners significantly outperformed control learners). Kamykowska et al. (2014) Standardised tests of phonics and letter-sound knowledge, word reading speed, pseudoword reading speed, non-verbal abilities, and receptive vocabulary.
No differences between the group who received GraphoGame intervention and a Maths intervention on study outcomes; no intervention effect.
Larabee et al.
Informal measures of phonics and letter-sound knowledge, words and pseudowords.
No difference between iPad instruction and instruction with typical materials on letter knowledge and word reading. Impact of iPad intervention on reading performance was mixed.