A recently published study from Stanford SCALE Initiative examines why one-on-one early literacy tutoring produces substantially larger learning gains than two-on-one tutoring, drawing on detailed transcript data from a large randomized controlled trial of virtual tutoring for K–2 students. Prior work from the same intervention found that one-on-one tutoring nearly doubled impacts on literacy outcomes compared to tutoring two students at a time. This study focuses on uncovering the mechanisms behind that difference.
Using natural language processing and machine learning methods, the authors analyze over 16,000 tutoring sessions and millions of tutor utterances to compare instructional practices, attention allocation, and relationship-building across formats. They examine session structure, how tutor time is spent, and the degree of personalization in instruction and interactions.
The results show that structural features—such as session length, frequency, and tutor consistency—are largely similar across formats. However, the distribution of tutor attention differs sharply. In one-on-one sessions, tutors devote nearly all instructional and relational interactions to a single student. In two-on-one sessions, attention is frequently split or shared, substantially reducing the amount of individualized instruction and relationship-building each student receives.
The analysis also reveals that one-on-one tutoring supports richer personalization, including greater responsiveness to students’ academic needs, increased use of encouragement and rapport-building talk, and more frequent leveraging of students’ linguistic and cultural backgrounds. Activities such as games are more affective in one-on-one settings, as peer interactions might create more tension or competition, while adult–student interactions tend to encourage trust and engagement.
The authors conclude that the power of one-on-one tutoring lies in undivided, individualized attention, not merely increased instructional time. They argue that small-group tutoring may close the effectiveness gap if it adopts intentional strategies to structure peer interaction, manage divided attention, and support tutors with targeted training or technology. These findings have direct implications for scaling tutoring programs cost-effectively without sacrificing impact.
