No abstract

Background The aims of this systematic review and meta-analyses were to determine if there is a statistically reliable association between poor reading and poor self-concept, and if such an association is moderated by domain of self-concept, type of reading impairment, or contextual factors including age, gender, reading instruction, and school environment. Methodology We searched 10 key databases for published and unpublished studies, as well as reference lists of included studies, and studies that cited included studies. We calculated standardised mean differences (SMDs) and 95% confidence intervals for one primary outcome (average self-concept) and 10 secondary outcomes (10 domains of self-concept). We assessed the data for risk of bias, heterogeneity, sensitivity, reporting bias, and quality of evidence. Results Thirteen studies with 3,348 participants met our selection criteria. Meta-analyses revealed statistically significant SMDs for average self-concept (−0.57) and five domains of self-concept (reading/writing/spelling: −1.03; academic: −0.67; math: −0.64; behaviour: −0.32; physical appearance: −0.28). The quality of evidence for the primary outcome was moderate, and for secondary outcomes was low, due to lack of data. Conclusions These outcomes suggest a probable moderate association between poor reading and average self-concept; a possible strong association between poor reading and reading-writing-spelling self-concept; and possible moderate associations between poor reading and self-concept in the self-concept domains of academia, mathematics, behaviour, and physical appearance.

Analysis 1.2. Comparison 1 Phonics training versus control (random-e ects model), Outcome 2 Non-word reading accuracy...... Analysis 1.3. Comparison 1 Phonics training versus control (random-e ects model), Outcome 3 Irregular word reading accuracy.