Dyslexia, affecting 5-17% of the population, has significant implications for child development. While research has shifted towards preventive approaches, current early screening tools often lack comprehensive language risk measures and scientific validation. Furthermore, it is unclear how well these language-related risk factors maintain their predictive power across reading development, highlighting a critical gap in our ability to predict dyslexia accurately, particularly for late-emerging cases (30-40% of cases) that often appear around fourth grade. Moreover, while most dyslexia studies focus on Western languages, language-specific features may significantly impact reading development and dyslexia manifestation, necessitating examination in diverse linguistic contexts.

This study investigated how well kindergarten language-related risk factors predict dyslexia in initial (first grade) and more advanced (fourth grade) stages of reading development in Hebrew, with its unique orthographic and morphological features. We addressed two key questions:
1. How do deficits in kindergarten language-related factors predict dyslexia in Grades 1 and 4?
2. What is the relative impact of deficits in each kindergarten language-related factor on the likelihood of dyslexia in Grades 1 and 4?

We conducted a 5-year longitudinal study following 515 Hebrew-speaking children (230 boys, 285 girls) from kindergarten through fourth grade. In kindergarten (mean age 5.9 years), children were assessed on phonological awareness (PA), rapid automatized naming (RAN), letter knowledge (LK), and morphological awareness (MA) abilities. Those scoring below the 16th percentile in these language-related measures were classified as at-risk. In first grade (mean age 6.8 years) and fourth grade (mean age 9.7 years), we identified disabled readers based on word reading fluency below the 16th percentile.

Logistic regression analysis revealed that deficits in early PA and LK were significant predictors of dyslexia in first grade, increasing likelihood by 4.13 (p<.001) and 2.92 (p<.001) times, respectively. By fourth grade, beyond first-grade dyslexia status (Exp(β)=7.11, p<.001), deficits in early MA and LK continued to exert significant direct influence, increasing dyslexia likelihood by 3.52 (p<.001) and 2.55 (p<.01) times, respectively.

These findings advance our understanding of dyslexia risk trajectories, demonstrating that the importance of language-related risk factors changes over time. The enduring impact of early LK deficits across grades underscores its critical role in dyslexia prediction throughout development. The emergence of early MA deficits as a significant predictor in fourth grade dyslexia highlights its importance in advanced reading stages, particularly in Hebrew's rich morphological system. This emphasizes the need for a multi-component early screening approach to dyslexia prediction and intervention.

Keywords: Dyslexia, Reading Development, Screening tools, Education.