Instructional Objectives Completed And Skill-Specific Early Reading Growth In A Digital Reading Tutor

Evaluations of digital and artificial-intelligence (AI)-supported reading tutors typically measure dose as time on the platform, yet minutes logged may be a poor proxy for the instruction studentsactually receive. In this study, we examined whether the number of instructional objectives students completed in the Project Read Tutor, a digital phonics and fluency tutor aligned to the University of Florida Literacy Institute Foundations scope and sequence, predicted early reading growth, and how that content-based dose compared with minutes of use. Participants were 811 kindergarten through Grade 2 tutor users in 43 classrooms in four Indiana elementary schools; outcomes were end-of-year DIBELS 8 national percentiles adjusted for beginning-of-year scores with classroom fixed effects. Each additional 10 objectives predicted 3.15 composite percentile points (g = 0.10). When both exposure measures were modeled jointly, objectives completed remained a statistically significant predictor (g = 0.19 per SD), whereas minutes were attenuated to nonsignificance (g = 0.09, p = .070). Gains were not detectable below approximately 10 objectives and rose thereafter. Mode-specific associations tracked instructional content: word-mode objectives were associated with code-focused skills and not with connected-text fluency, whereas story-mode objectives were associated with both, most strongly with code-focused skills. Associations were steeper for students who began the year with weaker skills and for students with disabilities. Because the tutor advances students on mastery, objectives completed partly reflect learning itself, and the associations are not causal. Overall, content completed was the stronger predictor of growth when both exposure measures were modeled together.