Educator Attention: How computational tools can systematically identify the distribution of a key resource for students

TL;DR

This study addresses how educators distribute attention across students in scalable tutoring settings and whether patterns align with equity goals. Leveraging over $1{,}157{,}970$ utterances from a large randomized tutoring program, the authors develop a two-dimensional attention framework (Recipient of Attention; Nature of Attention) and train a RoBERTa-based classifier to label utterances at scale. They find that while educators tend to devote more attention to lower-achieving students, gender-, race-, and EL-status demographics shape attention in meaningful ways, including a notable gender gap where lower-achieving girls in mixed-gender groups receive less attention than their male peers, and mixed-race/EL pairings showing differential attention patterns. The work demonstrates the value of large-scale observational data and NLP methods to reveal subtle disparities in instructional practice, informing more equitable tutoring strategies and future research on bidirectional student–teacher interactions.

Abstract

Educator attention is critical for student success, yet how educators distribute their attention across students remains poorly understood due to data and methodological constraints. This study presents the first large-scale computational analysis of educator attention patterns, leveraging over 1 million educator utterances from virtual group tutoring sessions linked to detailed student demographic and academic achievement data. Using natural language processing techniques, we systematically examine the recipient and nature of educator attention. Our findings reveal that educators often provide more attention to lower-achieving students. However, disparities emerge across demographic lines, particularly by gender. Girls tend to receive less attention when paired with boys, even when they are the lower achieving student in the group. Lower-achieving female students in mixed-gender pairs receive significantly less attention than their higher-achieving male peers, while lower-achieving male students receive significantly and substantially more attention than their higher-achieving female peers. We also find some differences by race and English learner (EL) status, with low-achieving Black students receiving additional attention only when paired with another Black student but not when paired with a non-Black peer. In contrast, higher-achieving EL students receive disproportionately more attention than their lower-achieving EL peers. This work highlights how large-scale interaction data and computational methods can uncover subtle but meaningful disparities in teaching practices, providing empirical insights to inform more equitable and effective educational strategies.

Figures (10)

• Figure 1: Individual attention allocation based on student achievement. (a) Lower-achieving students (L) receive significantly more individual attention than their higher-achieving peers (H). (b) Breakdown of attention by type, showing that lower-achieving students receive more attention on content-related support, relationship-building interactions, and session management. Asterisks indicate statistical significance (*** $p < 0.001$, ** $p < 0.01$).
• Figure 2: Regression coefficients for individual attention allocation across student pairings. The left-most dot is the reference group. Error bars indicate standard errors. (a) Gender pairings: In mixed-gender groups, female students receive significantly less attention than their male peers (the reference group). (b) Race pairings: In groups with different races with a focus on Black students, Black students receive slightly more attention than their non-Black peer (the reference group). (c) EL status pairings: In groups with students of different EL status, we find that students get about the same amount of attention.
• Figure 3: Distribution of individual attention by achievement across student pairings. Error bars indicate standard errors. (a) Gender pairings: In mixed-gender groups, lower-achieving male students receive significantly more attention than their female counterparts. (b) Race pairings: Lower-achieving Black students in Black-only groups receive significantly more attention, but this pattern does not hold in mixed-race pairings. (c) English Learner (EL) status pairings: Lower-achieving EL students receive less attention when paired with another EL student. The lower-achieving non-EL student receives more attention in all settings.
• Figure 4: Nature of attention by gender.
• Figure 5: Nature of attention by race.