A l ongitudinal a nalysis of AI-based basic literacy development: Focusing on early decoding and reading fluency indicators
This study examined the longitudinal relationship between Grade 3 early decoding and reading fluency indicators and Grade 5 AI-based English basic literacy diagnostic results. Data were collected from 50 elementary school students with both Grade 3 and Grade 5 diagnostic data. Grade 3 indicators included Correct Letter Sounds and Whole Words Read from Nonsense Word Fluency, as well as Oral Reading Fluency reading speed and accuracy rate. Grade 5 results included phonemic awareness, basic phonics, reading fluency and comprehension, and overall basic literacy performance. Descriptive statistics, one-way ANOVA, correlation analysis, and simple regression analysis were conducted. Results showed that Grade 5 performance was generally at the middle or upper level, while basic phonics was relatively weaker. Significant differences by early decoding level appeared in reading fluency and comprehension and overall basic literacy performance. All Grade 3 indicators were significantly correlated with Grade 5 basic literacy performance, with reading accuracy rate showing the strongest association. These findings suggest that early decoding and oral reading fluency indicators can inform later English literacy diagnosis and instruction.