Digital Tools for Reading Success: Meta-analyses of Digital Interventions

BackgroundGiven the growing reliance on digital devices, an increasing number of studies have examined the effects of text medium on reading outcomes in development; however, the results have been mixed. The goal of this systematic review is to look at how print and digital formats affect reading comprehension, engagement and other reading outcomes (e.g. vocabulary, reading speed) in children and adolescents aged 1–17 years old while also considering the influence of several participant, task and study characteristics.MethodsA comprehensive search strategy involving seven electronic databases yielded 88 eligible articles comparing digital and print formats on reading outcomes published between 2000 and 2023 (3 reviewer inter‐rater reliability: k = .54–.78). Three major characteristics were coded: participant‐level (grade/age, diverse populations, testing language); task‐level (text‐genre, shared reading, digital comparability); study‐level (publication recency, study quality) characteristics. Contingency tables were created for all studies, then for each reading outcome and for participant, task, and study characteristics separately to classify the percentage of studies that demonstrated outcomes favouring print, digital, no difference or reliance on specific reading measures or other factors.ResultsExcept in the case of engagement as an outcome, the most common finding was no difference between digital and print. When participant, task and study characteristics were examined separately for the various reading outcomes, the results varied. More studies examining reading comprehension (particularly of informational text and in older children) found ‘print is better’, whereas ‘digital is better’ was more common in studies examining engagement, other outcomes such as vocabulary and diverse learners.ConclusionsThis review highlights the importance of examining multiple interacting factors when studying the impact of print versus digital mediums on reading outcomes in children and adolescents.

Previous studies have shown that higher socioeconomic status (SES) and richer home literacy environment (HLE) are associated with better reading outcomes in children with family risk for reading difficulties (RD). Yet, the cognitive and neural mechanisms underlying this association remain understudied. This study sought to fill in the gap using both behavioral and neuroimaging methodologies. Eighty-one native Mandarin-speaking children (age = 8.7-12.0 years) and their parents were recruited. Family history (FH) of RD and home environmental factors (SES and HLE) were collected, as well as children's diffusion MRI, reading-related cognitive skills (rapid automatized naming [RAN], phonological awareness [PA], and morphological awareness [MA]), and reading outcomes. Participants were divided into the following groups according to their FH of RD and current reading status: children with a FH of RD who showed typical development (FH+TD) or who developed RD (FH+RD), and a typically developing control group without a FH (FH-TD). Chain mediation analyses showed that in FH+TD, the number of children's books (NCB) was linked to the axial diffusivity (AxD) of children's right superior longitudinal fasciculus (rSLF). This neural metric was further linked to children's silent reading comprehension (SRC) via their MA. No such mediation relationship was shown in either FH+RD or FH-TD. These results suggest that reading-related cognitive skills constitute important pathways linking HLE and reading outcomes in at-risk children who developed typical reading skills, while the right-hemispheric white matter fiber tract may serve as one possible neural intermediary. Our findings have educational implications for developing early interventions that focus on enriching HLE to help mitigate the risk of RD. SUMMARY: Number of children's books (NCB) is uniquely associated with reading development in children with a family history (FH) of reading difficulties (RD) who nevertheless develop typical reading abilities. Rapid automatized naming (RAN) serves as a significant cognitive mediator for oral reading fluency (ORF), and morphological awareness (MA) for silent reading comprehension (SRC). One possible neural mechanism underlying the NCB → MA → SRC mediation pathway could be the white matter structure, rSLF-AxD. An updated operational definition is proposed and applied to identify protective factors in children at risk for RD. The chain mediation method is used to examine the multi-level pathways (environment-brain-cognition-behavior), as hypothesized by the Multiple Deficit Model (MDM) of reading development.

Serial order recall for visually and auditorily presented stimuli was examined in a group of 12-year-old poor readers and 7-year-old reading-age controls. With pictorial presentation, the poor readers showed a visual similarity effect, no word length effect, and a smaller phonemic similarity effect than that of controls. However, with visual presentation of printed words and with auditory presentation, poor readers showed word length and phonemic similarity effects of similar magnitude to that of controls. It is concluded that poor readers rely on visual information in tasks where the presented images are highly codable, and where verbal recoding is not obligatory, but that they will make use of phonological coding when the stimuli are not as easily codable visually in memory.

ABSTRACTThis Study investigated whether the pseudoword repetition difference between poor and normal readers could be explained by differences in memory for verbal materials or in familiarity with the composition of verbal materials. Dutch second graders and poor readers scoring in the same range on a word-reading test repeated pseudowords that varied in length and wordlikeness. The pseudoword repetition deficit of poor readers reported in the literature was replicated. Although the repetition scores were influenced by pseudoword length and wordlikeness, no interaction with reader group was found, thus indicating that neither variable represented a causal factor in the poor readers' repetition deficit. Statistical correction for reader group differences on a phoneme detection task and an auditory discrimination task made the pseudoword repetition difference disappear. We conclude that the basic problem is unlikely to be with subvocal rehearsal, and that the cause of the repetition deficit is already operative in early, perceptual stages of processing.

IntroductionThis study compared the phonological, musical, and general cognitive profiles of early-emerging poor, average, and good readers.MethodsWe assessed Cantonese preschool children on Chinese word reading, phonological awareness, lexical tone awareness, musical rhythm perception, musical pitch perception, working memory, and non-verbal intelligence.ResultsEarly-emerging poor readers exhibited poorer phonological awareness than early-emerging average and good readers, whereas the latter two groups did not differ significantly. In the working memory task, early-emerging good readers outperformed both early-emerging average and poor readers, who performed similarly. No significant group differences were found in lexical tone awareness, musical rhythm perception, musical pitch perception, or non-verbal intelligence.DiscussionThe results reflect phonological deficits in early-emerging poor readers. Furthermore, phonological awareness and working memory were useful for identifying early-emerging poor and good readers, respectively. Clinically, these findings imply that early-emerging poor readers may benefit most from initial phonological awareness training, followed by working memory training. Moreover, working memory training may also be beneficial for early-emerging average readers seeking to improve their Chinese word reading.

This paper examines the relationships among the quality and quantity of parent–child shared book reading (SBR) engagements and children's reading and mathematics outcomes in preschool. Additionally, we explore how child and family characteristics predict the quality and quantity of SBR. Quantity was measured using parental reports of the frequency of SBR. Quality was measured by observational protocols evaluating for questioning, vocabulary, and discussion depth. A structural equation model was estimated using data from a nationally representative sample of 700 children living in the United States from the Early Childhood Longitudinal Study—Birth cohort. Results indicated that the quality of book reading was associated with children's mathematics outcomes, and the quantity was associated with reading outcomes controlling for contextual variables. Socioeconomic status, race or ethnicity, and children's age predicted the quality and quantity of book reading experiences. These findings indicate that frequent exposure to high‐quality book reading may positively impact children's mathematics and reading development, but that variation in SBR exists. SBR is a common practice among many parents; helping parents understand the multiple benefits of the practice may further increase the frequency and quality of the engagement. Implications for research and practice are addressed.Highlights This paper describes the relationship between parental book reading practices, mathematics, and reading outcomes. Structural equation models reveal relationships between the quantity of book reading and children's reading outcomes, as well as the quality of book reading and children's mathematics outcomes. High‐quality book reading may positively impact academic achievement in multiple domains, but results may vary based on contextual factors.

The purpose of this study was to investigate literacy development beyond the early school years. We investigated mean-level and rank-order changes in two reading (word decoding speed and reading comprehension) and two spelling (word and pseudoword spelling accuracy) tasks during a 4-year period from beginning to the later grades of elementary school, and examined whether rank-order changes in literacy skills could be predicted from verbal working memory, phonological awareness, and morpho-syntactic awareness. The sample consisted of Croatian speaking children, poor (n = 50) and good (n = 61) readers. Croatian is a language with a highly transparent orthography but a rather complex grammar. Word decoding speed and word spelling accuracy improved significantly over the study period, word decoding speed more so among good than among poor readers, supporting the Matthew effect hypothesis. Literacy skills were moderately stable, with the highest stability coefficients obtained for word decoding speed, and the lowest for pseudoword spelling. The predictors of rank-order changes varied across literacy outcomes, but did not differ for poor and good readers. Morpho-syntactic awareness predicted rank-order development of all outcomes except for word spelling. Phonological awareness predicted rank-order development of reading comprehension and word spelling. Verbal working memory predicted rank-order development of word spelling only, and at a low level. The finding that the more language related cognitive variables, i.e., morpho-syntactic and phonological skills, were stronger predictors of literacy development than working memory, for both poor and good readers, suggests remedial focus on these more predictive variables.

Deficits in verbal short-term memory have been identified as one factor underlying reading and spelling disorders. However, the nature of this deficit is still unclear. It has been proposed that poor readers make less use of phonological coding, especially if the task can be solved through visual strategies. In the framework of Baddeley's phonological loop model, this study examined serial recall performance in German second-grade children with poor vs good reading and spelling abilities. Children were presented with four-item lists of common nouns for immediate serial recall. Word length and phonological similarity as well as presentation modality (visual vs auditory) and type of recall (visual vs verbal) were varied as within-subject factors in a mixed design. Word length and phonological similarity effects did not differ between groups, thus indicating equal use of phonological coding and rehearsal in poor and good readers. However, in all conditions, except the one that combined visual presentation and visual recall, overall performance was significantly lower in poor readers. The results suggest that the poor readers' difficulties do not arise from an avoidance of the phonological loop, but from its inefficient use. An alternative account referring to unstable phonological representations in long-term memory is discussed.

Working memory (WM) training in typically developing (TD) children aims to enhance not only performance in memory tasks but also other domain-general cognitive skills, such as fluid intelligence. These benefits are then believed to positively affect academic achievement. Despite the numerous studies carried out, researchers still disagree over the real benefits of WM training. With this meta-analysis (m = 41, k = 393, N = 2,375), we intended to resolve the discrepancies by focusing on the potential sources of within-study and between-study true heterogeneity. Small to medium effects were observed in memory tasks (i.e., near transfer). The size of these effects was proportional to the similarity between the training task and the outcome measure. By contrast, far-transfer measures of cognitive ability (e.g., intelligence) and academic achievement (mathematics and language ability) were essentially unaffected by the training programs, especially when the studies implemented active controls ([Formula: see text] = 0.001, SE = 0.055, p = .982, τ2 = 0.000). Crucially, all the models exhibited a null or low amount of true heterogeneity, which was wholly explained by the type of controls (nonactive vs. active) and by statistical artifacts, in contrast to the claim that this field has produced mixed results. Since the empirical evidence shows the absence of both generalized effects and true heterogeneity, we conclude that there is no reason to keep investing resources in WM training research with TD children.

This paper reports the results of two studies of reading comprehension of Flemish children in Belgium. In the northern part of Belgium (Flanders), Dutch is the official language. The Dutch-speaking inhabitants of Flanders are called Flemish. Dutch is also the national language of the Netherlands. Despite both groups using Dutch, cultural differences between the groups are reflected in healthcare and education. Study 1 investigated the effect of these differences on the reading comprehension of deaf children. Reading levels of Flemish deaf children with cochlear implants (CI) (n = 30) or hearing aids (HA) (n = 44) were compared with reference data of Dutch deaf children with CI (n = 50) and HA (n = 500), and with a hearing norm group. Study 2 investigated sources of variability by examining the underlying processes of reading comprehension of good and poor Flemish deaf readers from the CI group in Study 1. Results of Study 1 showed significantly better reading levels for the two Flemish groups. These differences may reflect Belgian policy aiming at early implantation and the use of spoken language communication. The second study contrasted good and poor readers' working memory capacity scores, including verbal and non-verbal working memory scores, morphosyntactic ability, and phonological encoding in order to explore which underlying processes contributed to reading performance in the Flemish CI users. Results showed a tendency toward better morphosyntactic abilities and better working memory skills in the good readers. Three factors appear to explain the better reading of Flemish children with CIs: amount of spoken language input in communication, access to spoken language through a cochlear implant, and age at implantation.

Sensitivity to print is characterized by a left occipito-temporal negativity to words in the event-related potential N1. This sensitivity is modulated by reading skills and may thus represent a neural marker of reading competence. Here we studied the development of the N1 in regular and poor readers from preschool age to school age to test whether the amplitude of the N1 predicts children's reading outcomes. Our results suggest a predictive value of the print-sensitive negativity over the right hemisphere. Whether this N1 may serve as a biomarker to improve prognosis in preliterate children should be clarified in future studies.

This study reports on several specific neurocognitive processes and eye-tracking predictors of reading outcomes for a sample of children with Developmental Dyslexia (DD) and At-tention-Deficit/Hyperactivity Disorder - inattentive subtype (ADHD-I) compared to typical readers. Participants included 19 typical readers, 21 children diagnosed with ADHD-I and 19 children with DD. All participants were attending 4th grade and had a mean age of 9.08 years. The psycholinguistic profile of each group was assessed using a battery of neuropsy-chological and linguistic tests. Participants were submitted to a silent reading task with lex-ical manipulation of the text. Multinomial logistic regression was conducted to evaluate the predictive capability of developing dyslexia or ADHD-I based on the following measures: (a) a linguistic model that included measures of phonological awareness, rapid naming, and reading fluency and accuracy; (b) a cognitive neuropsychological model that included measures of memory, attention, visual processes, and cognitive or intellectual functioning, and (c) an additive model of lexical word properties with manipulation of word-frequency and word-length effects through eye-tracking. The additive model in conjunction with the neuropsychological model classification improved the prediction of who develops dyslexia or ADHD-I having as baseline normal readers. Several of the neuropsychological and eye-tracking variables have power to predict the degree of reading outcomes in children with learning disabilities.

Many children with Developmental Language Disorder (DLD) develop reading difficulties. The purpose of this study is to better understand variation in the reading outcomes of children with DLD using a person-centered approach. 87 monolingual Dutch children diagnosed with DLD performed at ages 5 or 6 years nine tests of nonverbal IQ, oral language proficiency, phonological memory (PM) and executive functioning (EF). Two years later, the same children were tested on single (non-)word reading. Latent profile analyses were conducted to identify profiles based on oral language proficiency, phonological memory and executive functioning at age 5-6 years, which, in turn, were related to nonverbal IQ and to single-word reading two years later. Four profiles were identified and labelled relative to their position within the DLD-sample: 1. Weak performance overall, 2. Strong EF-average language and PM, 3. Mild working memory (WM) deficiencies-average language and PM, 4. Strong development overall. Profiles 1 and 3 had below average nonverbal IQ scores and were associated with low word reading outcomes two years later. Within the group of children with DLD, children with relatively weak oral language, phonological memory and executive functioning, or children with working memory deficiencies are most at risk for developing reading difficulties. The findings support a multiple risk framework and confirm that a person-centered approach is promising in predicting reading outcomes in DLD.Implications: Research into individual differences in DLD is dominated by variable-centered approaches. This study illustrates how a person-centered approach, which views variables as properties of individuals, captures variation in the DLD-population. Using this bottom-up approach, the study highlights how an individual's strengths and weaknesses across different developmental domains can be combined into profiles that relate to later reading outcomes. As such, it can provide an example for future DLD research.

Single-case experimental designs (SCEDs) are used to study the effects of interventions on the behavior of individual cases, by making comparisons between repeated measurements of an outcome under different conditions. In research areas where SCEDs are prevalent, there is a need for methods to synthesize results across multiple studies. One approach to synthesis uses a multilevel meta-analysis (MLMA) model to describe the distribution of effect sizes across studies and across cases within studies. However, MLMA relies on having accurate sampling variances of effect size estimates for each case, which may not be possible due to auto-correlation in the raw data series. One possible solution is to combine MLMA with robust variance estimation (RVE), which provides valid assessments of uncertainty even if the sampling variances of effect size estimates are inaccurate. Another possible solution is to forgo MLMA and use simpler, ordinary least squares (OLS) methods with RVE. This study evaluates the performance of effect size estimators and methods of synthesizing SCEDs in the presence of auto-correlation, for several different effect size metrics, via a Monte Carlo simulation designed to emulate the features of real data series. Results demonstrate that the MLMA model with RVE performs properly in terms of bias, accuracy, and confidence interval coverage for estimating overall average log response ratios. The OLS estimator corrected with RVE performs the best in estimating overall average Tau effect sizes. None of the available methods perform adequately for meta-analysis of within-case standardized mean differences. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

In this preregistered multi-level meta-analysis, we aim to clarify the association of need for cognition (NFC) and typical intellectual engagement (TIE) with intelligence and executive functions. Multi-level models with robust variance estimation were specified and risk of bias was assessed with the adapted Risk of Bias Utilized for Surveys Tool. NFC/TIE was associated with fluid intelligence (Gf; r = 0.18, p < .001, N = 25,367), crystallized intelligence (Gc; r = 0.26, p < .001, N = 14,651), general intelligence (r = 0.23, p < .001, N = 8479), and working memory (WM) capacity (r = 0.15, p < .001, N = 5921). Associations with WM updating (r = 0.08, p = .111, N = 1084), inhibition (r = 0.04, p = .077, N = 2895), and shifting (r = 0.01, p = 0.642, N = 1727) were non-significant. NFC (r = 0.19) was more strongly related to Gf than TIE (r = 0.12; F(1, 12.10) = 5.04, p = .045) whereas TIE (r = 0.35) was more strongly associated with Gc than NFC (r = 0.24; F(1, 13.10) = 10.70, p = .006). Correlations with Gc significantly declined over time (b1 = −0.006, β1 = −0.04, p = .010). Results provide strong evidence for small-to-moderate associations of NFC and TIE with Gf, Gc, general intelligence, and WM capacity, and at best small associations with core executive functions.