Script-Specific Predictors of Reading Comprehension in Hebrew-Speaking Children With Reading Difficulties

This study investigates cognitive-linguistic predictors of reading comprehension (RC) in Hebrew-speaking children with reading difficulties (RD). Hebrew's dual orthographic system including vowelized (shallow) and unvowelized (deep) scripts offers a valuable context for examining how cognitive-linguistic predictors differentially support reading across levels of orthographic transparency. A cross-sectional design was used with Grade 3 children with RD, utilising Fisher's Z-tests to identify significant differences in the strength of correlations between predictors and RC across scripts as well as multilevel modelling (MLM) to clarify the unique contribution of each predictor within each orthographic context. Results indicate that PA showed a stronger contribution to RC in the vowelized script, whereas RAN, vocabulary, and unvowelized word reading efficiency were stronger predictors in the unvowelized script. Cognitive flexibility was associated with RC in both scripts based on Pearson correlations, but MLM indicated a significant contribution only in the unvowelized script. These findings highlight the importance of considering orthographic depth when evaluating the cognitive-linguistic underpinnings of reading and offer script-specific insights that may inform assessment practices and intervention planning for children with RD.