Jing Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jing Liu has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Carol Espy-Wilson
Department of Electrical and Computer Engineering University of Maryland College Park USA
4 shared publications - Dorottya Demszky
Graduate School of Education, Stanford University, Stanford, USA
4 shared publications - Ahmed Adel Attia
University of Maryland College Park MD
3 shared publications - Wei Ai
University of Maryland College Park
3 shared publications - Francisco Lagos
Inter-American Development Bank, Washington, DC
2 shared publications - Heather C. Hill
Harvard University, Cambridge, MA
2 shared publications
Latest Publications
- Open Access
- 2 October 2026
- Published by Elsevier in SSRN Electronic Journal
The Effects of Course-integrated AI Tutoring on Student Performance and Engagement: A Randomized University Trial
We evaluate access to a course-integrated generative AI (GenAI) tutor in a randomized controlled trial involving 2,379 undergraduate students and 30 instructors across multiple disciplines in a large U.S. public university. Among sections of the same course, tutor access reduced final grades by 0.37 standard deviations and learning management system participation by 0.90 standard deviations. Participation declines were similarly large in the full sample, and estimated academic losses were greater among first-generation students. These findings suggest that introducing an AI tutor may displace engagement with existing learning activities without improving academic performance, with potentially differential impact across student groups.
- 2 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessThe Effects of Course-integrated AI Tutoring on Student Performance and Engagement: A Randomized University Trial
We evaluate access to a course-integrated generative AI (GenAI) tutor in a randomized controlled trial involving 2,379 undergraduate students and 30 instructors across multiple disciplines in a large U.S. public university. Among sections of the same course, tutor access reduced final grades by 0.37 standard deviations and learning management system participation by 0.90 standard deviations. Participation declines were similarly large in the full sample, and estimated academic losses were greater among first-generation students. These findings suggest that introducing an AI tutor may displace engagement with existing learning activities without improving academic performance, with potentially differential impact across student groups.
- Open Access
- 2 September 2026
- Published by Elsevier in SSRN Electronic Journal
The Effect of Centralized School Assignment Systems on Between-School Segregation: Evidence from the 300 Largest School Districts in the United States
This study examines how the adoption of centralized school assignment systems affects between-school racial and ethnic segregation in the largest U.S. public school districts. Using an original nationwide panel dataset and a staggered difference-in-differences design, we estimate dynamic and average changes in district composition, intergroup exposure, and the evenness of student distributions across schools. Event-study estimates show that the main compositional and exposure changes generally emerge two to three years after adoption rather than as an immediate level shift. Adoption is associated on average with a 2.0 percentage point increase in the White share of district enrollment and modest, asymmetric changes in exposure: Black-White exposure increases by 1.6 percentage points and student of color-White exposure by 1.8 points, while White students’ exposure to Black, Asian, and students of color declines. These patterns can coexist because exposure is group-specific and depends on both group size and the schools attended by each group. We find no statistically significant effect on dissimilarity at any point in the event window. The results therefore indicate that, on average, moving to these types of centralized assignment systems produced gradual but modest changes in composition and peer exposure without materially altering the evenness of racial and ethnic sorting across schools.
- 2 September 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessThe Effect of Centralized School Assignment Systems on Between-School Segregation: Evidence from the 300 Largest School Districts in the United States
This study examines how the adoption of centralized school assignment systems affects between-school racial and ethnic segregation in the largest U.S. public school districts. Using an original nationwide panel dataset and a staggered difference-in-differences design, we estimate dynamic and average changes in district composition, intergroup exposure, and the evenness of student distributions across schools. Event-study estimates show that the main compositional and exposure changes generally emerge two to three years after adoption rather than as an immediate level shift. Adoption is associated on average with a 2.0 percentage point increase in the White share of district enrollment and modest, asymmetric changes in exposure: Black-White exposure increases by 1.6 percentage points and student of color-White exposure by 1.8 points, while White students’ exposure to Black, Asian, and students of color declines. These patterns can coexist because exposure is group-specific and depends on both group size and the schools attended by each group. We find no statistically significant effect on dissimilarity at any point in the event window. The results therefore indicate that, on average, moving to these types of centralized assignment systems produced gradual but modest changes in composition and peer exposure without materially altering the evenness of racial and ethnic sorting across schools.
- Open Access
- 9 July 2026
- Published by Elsevier in SSRN Electronic Journal
JUE Insight: The Effect of Centralized School Assignment Systems on Between-School Segregation: Evidence from the 300 Largest School Districts in the United States
This study examines how the adoption of centralized school assignment systems affects between-school racial and ethnic segregation in the largest U.S. public school districts. Using an original nationwide panel dataset and a staggered difference-in-differences design, we estimate dynamic and average changes in district composition, intergroup exposure, and the evenness of student distributions across schools. Event-study estimates show that the main compositional and exposure changes generally emerge two to three years after adoption rather than as an immediate level shift. Adoption is associated on average with a 2.0 percentage point increase in the White share of district enrollment and modest, asymmetric changes in exposure: Black-White exposure increases by 1.6 percentage points and student of color-White exposure by 1.8 points, while White students’ exposure to Black, Asian, and students of color declines. These patterns can coexist because exposure is group-specific and depends on both group size and the schools attended by each group. We find no statistically significant effect on dissimilarity at any point in the event window. The results therefore indicate that, on average, moving to these types of centralized assignment systems produced gradual but modest changes in composition and peer exposure without materially altering the evenness of racial and ethnic sorting across schools.
- 9 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessJUE Insight: The Effect of Centralized School Assignment Systems on Between-School Segregation: Evidence from the 300 Largest School Districts in the United States
This study examines how the adoption of centralized school assignment systems affects between-school racial and ethnic segregation in the largest U.S. public school districts. Using an original nationwide panel dataset and a staggered difference-in-differences design, we estimate dynamic and average changes in district composition, intergroup exposure, and the evenness of student distributions across schools. Event-study estimates show that the main compositional and exposure changes generally emerge two to three years after adoption rather than as an immediate level shift. Adoption is associated on average with a 2.0 percentage point increase in the White share of district enrollment and modest, asymmetric changes in exposure: Black-White exposure increases by 1.6 percentage points and student of color-White exposure by 1.8 points, while White students’ exposure to Black, Asian, and students of color declines. These patterns can coexist because exposure is group-specific and depends on both group size and the schools attended by each group. We find no statistically significant effect on dissimilarity at any point in the event window. The results therefore indicate that, on average, moving to these types of centralized assignment systems produced gradual but modest changes in composition and peer exposure without materially altering the evenness of racial and ethnic sorting across schools.
- 30 November 2025
- Published by American Educational Research Association (AERA) in Educational Researcher
- Vol. 55 (4) , 255-265
A Quantitative Study of Mathematical Vocabulary Use in Upper Elementary Classrooms
This study provides the first large-scale quantitative exploration of mathematical vocabulary use in upper elementary U.S. classrooms. Our approach employs natural language processing techniques to describe variation in teachers’ and students’ use of mathematical vocabulary in 1,657 fourth- and fifth-grade lessons in 317 classrooms in four districts over 3 years. Students’ exposure to mathematical vocabulary varies substantially across lessons and between teachers. Results suggest that teacher modeling, defined as the frequency of mathematical vocabulary in teacher talk, does not substantially cause students to uptake mathematical vocabulary but that teachers may encourage student use of mathematical vocabulary by means other than mere modeling or exposure. We find that teachers who more frequently model mathematical vocabulary use are more effective at raising student test scores.
- 30 November 2025
- Published by American Educational Research Association (AERA) in Educational Researcher
- Vol. 55 (4) , 255-265
A Quantitative Study of Mathematical Vocabulary Use in Upper Elementary Classrooms
This study provides the first large-scale quantitative exploration of mathematical vocabulary use in upper elementary U.S. classrooms. Our approach employs natural language processing techniques to describe variation in teachers’ and students’ use of mathematical vocabulary in 1,657 fourth- and fifth-grade lessons in 317 classrooms in four districts over 3 years. Students’ exposure to mathematical vocabulary varies substantially across lessons and between teachers. Results suggest that teacher modeling, defined as the frequency of mathematical vocabulary in teacher talk, does not substantially cause students to uptake mathematical vocabulary but that teachers may encourage student use of mathematical vocabulary by means other than mere modeling or exposure. We find that teachers who more frequently model mathematical vocabulary use are more effective at raising student test scores.
- Open Access
RealClass: A Framework for Classroom Speech Simulation with Public Datasets and Game Engines
The scarcity of large-scale classroom speech data has hindered the development of AI-driven speech models for education. Classroom datasets remain limited and not publicly available, and the absence of dedicated classroom noise or Room Impulse Response (RIR) corpora prevents the use of standard data augmentation techniques. In this paper, we introduce a scalable methodology for synthesizing classroom noise and RIRs using game engines, a versatile framework that can extend to other domains beyond the classroom. Building on this methodology, we present RealClass, a dataset that combines a synthesized classroom noise corpus with a classroom speech dataset compiled from publicly available corpora. The speech data pairs a children's speech corpus with instructional speech extracted from YouTube videos to approximate real classroom interactions in clean conditions. Experiments on clean and noisy speech show that RealClass closely approximates real classroom speech, making it a valuable asset in the absence of abundant real classroom speech.
Open AccessRealClass: A Framework for Classroom Speech Simulation with Public Datasets and Game Engines
The scarcity of large-scale classroom speech data has hindered the development of AI-driven speech models for education. Classroom datasets remain limited and not publicly available, and the absence of dedicated classroom noise or Room Impulse Response (RIR) corpora prevents the use of standard data augmentation techniques. In this paper, we introduce a scalable methodology for synthesizing classroom noise and RIRs using game engines, a versatile framework that can extend to other domains beyond the classroom. Building on this methodology, we present RealClass, a dataset that combines a synthesized classroom noise corpus with a classroom speech dataset compiled from publicly available corpora. The speech data pairs a children's speech corpus with instructional speech extracted from YouTube videos to approximate real classroom interactions in clean conditions. Experiments on clean and noisy speech show that RealClass closely approximates real classroom speech, making it a valuable asset in the absence of abundant real classroom speech.
- Open Access
SimClass: A Classroom Speech Dataset Generated via Game Engine Simulation For Automatic Speech Recognition Research
The scarcity of large-scale classroom speech data has hindered the development of AI-driven speech models for education. Public classroom datasets remain limited, and the lack of a dedicated classroom noise corpus prevents the use of standard data augmentation techniques. In this paper, we introduce a scalable methodology for synthesizing classroom noise using game engines, a framework that extends to other domains. Using this methodology, we present SimClass, a dataset that includes both a synthesized classroom noise corpus and a simulated classroom speech dataset. The speech data is generated by pairing a public children's speech corpus with YouTube lecture videos to approximate real classroom interactions in clean conditions. Our experiments on clean and noisy speech demonstrate that SimClass closely approximates real classroom speech, making it a valuable resource for developing robust speech recognition and enhancement models.
Open AccessSimClass: A Classroom Speech Dataset Generated via Game Engine Simulation For Automatic Speech Recognition Research
The scarcity of large-scale classroom speech data has hindered the development of AI-driven speech models for education. Public classroom datasets remain limited, and the lack of a dedicated classroom noise corpus prevents the use of standard data augmentation techniques. In this paper, we introduce a scalable methodology for synthesizing classroom noise using game engines, a framework that extends to other domains. Using this methodology, we present SimClass, a dataset that includes both a synthesized classroom noise corpus and a simulated classroom speech dataset. The speech data is generated by pairing a public children's speech corpus with YouTube lecture videos to approximate real classroom interactions in clean conditions. Our experiments on clean and noisy speech demonstrate that SimClass closely approximates real classroom speech, making it a valuable resource for developing robust speech recognition and enhancement models.
- Open Access
FT-Boosted SV: Towards Noise Robust Speaker Verification for English Speaking Classroom Environments
Creating Speaker Verification (SV) systems for classroom settings that are robust to classroom noises such as babble noise is crucial for the development of AI tools that assist educational environments. In this work, we study the efficacy of finetuning with augmented children datasets to adapt the x-vector and ECAPA-TDNN to classroom environments. We demonstrate that finetuning with augmented children's datasets is powerful in that regard and reduces the Equal Error Rate (EER) of x-vector and ECAPA-TDNN models for both classroom datasets and children speech datasets. Notably, this method reduces EER of the ECAPA-TDNN model on average by half (a 5 % improvement) for classrooms in the MPT dataset compared to the ECAPA-TDNN baseline model. The x-vector model shows an 8 % average improvement for classrooms in the NCTE dataset compared to its baseline.
Open AccessFT-Boosted SV: Towards Noise Robust Speaker Verification for English Speaking Classroom Environments
Creating Speaker Verification (SV) systems for classroom settings that are robust to classroom noises such as babble noise is crucial for the development of AI tools that assist educational environments. In this work, we study the efficacy of finetuning with augmented children datasets to adapt the x-vector and ECAPA-TDNN to classroom environments. We demonstrate that finetuning with augmented children's datasets is powerful in that regard and reduces the Equal Error Rate (EER) of x-vector and ECAPA-TDNN models for both classroom datasets and children speech datasets. Notably, this method reduces EER of the ECAPA-TDNN model on average by half (a 5 % improvement) for classrooms in the MPT dataset compared to the ECAPA-TDNN baseline model. The x-vector model shows an 8 % average improvement for classrooms in the NCTE dataset compared to its baseline.
- Open Access
- 1 January 2025
- Published by Elsevier in SSRN Electronic Journal
The Effect of Centralized-Admission School Lotteries on Between-School Segregation: Evidence from 300 Largest School Districts in the United States
This study examines how centralized-admission school lotteries affect between-school racial and ethnic segregation in the largest U.S. public school districts. Using original nationwide panel data and a difference-in-differences design with staggered adoption, the research analyzes effects on school composition, intergroup exposure, and distribution evenness. The findings reveal that centralized-admission lotteries led to increased White student enrollment in district schools and modest improvements in intergroup exposure. Black-White exposure rose by 1.6 percentage points and student of color-White exposure by 1.8 points. However, White students experienced reduced exposure to all racial and ethnic groups, with similar patterns for Black, Asian, and other students of color. While centralized lotteries modestly redistribute students, they do not significantly reduce overall segregation, challenging assumptions about equity-promoting reforms. These results underscore the need for complementary policies including weighted lottery designs, transportation subsidies, and targeted adoption to address the structural roots of school segregation.
- 1 January 2025
- Published by Elsevier in SSRN Electronic Journal
Open AccessThe Effect of Centralized-Admission School Lotteries on Between-School Segregation: Evidence from 300 Largest School Districts in the United States
This study examines how centralized-admission school lotteries affect between-school racial and ethnic segregation in the largest U.S. public school districts. Using original nationwide panel data and a difference-in-differences design with staggered adoption, the research analyzes effects on school composition, intergroup exposure, and distribution evenness. The findings reveal that centralized-admission lotteries led to increased White student enrollment in district schools and modest improvements in intergroup exposure. Black-White exposure rose by 1.6 percentage points and student of color-White exposure by 1.8 points. However, White students experienced reduced exposure to all racial and ethnic groups, with similar patterns for Black, Asian, and other students of color. While centralized lotteries modestly redistribute students, they do not significantly reduce overall segregation, challenging assumptions about equity-promoting reforms. These results underscore the need for complementary policies including weighted lottery designs, transportation subsidies, and targeted adoption to address the structural roots of school segregation.
- Open Access
- 16 October 2024
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society
- Vol. 7, 74-80
Kid-Whisper: Towards Bridging the Performance Gap in Automatic Speech Recognition for Children VS. Adults
Recent advancements in Automatic Speech Recognition (ASR) systems, exemplified by Whisper, have demonstrated the potential of these systems to approach human-level performance given sufficient data. However, this progress doesn’t readily extend to ASR for children due to the lim- ited availability of suitable child-specific databases and the distinct characteristics of children’s speech. A recent study investigated leveraging the My Science Tutor (MyST) chil- dren’s speech corpus to enhance Whisper’s performance in recognizing children’s speech. They were able to demon- strate some improvement on a limited testset. This paper builds on these findings by enhancing the utility of the MyST dataset through more efficient data preprocessing. We reduce the Word Error Rate (WER) on the MyST testset 13.93% to 9.11% with Whisper-Small and from 13.23% to 8.61% with Whisper-Medium and show that this improvement can be generalized to unseen datasets. We also highlight important challenges towards improving children’s ASR performance and the effect of fine-tuning in improving the transcription of disfluent speech.
- 16 October 2024
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society
- Vol. 7, 74-80
Open AccessKid-Whisper: Towards Bridging the Performance Gap in Automatic Speech Recognition for Children VS. Adults
Recent advancements in Automatic Speech Recognition (ASR) systems, exemplified by Whisper, have demonstrated the potential of these systems to approach human-level performance given sufficient data. However, this progress doesn’t readily extend to ASR for children due to the lim- ited availability of suitable child-specific databases and the distinct characteristics of children’s speech. A recent study investigated leveraging the My Science Tutor (MyST) chil- dren’s speech corpus to enhance Whisper’s performance in recognizing children’s speech. They were able to demon- strate some improvement on a limited testset. This paper builds on these findings by enhancing the utility of the MyST dataset through more efficient data preprocessing. We reduce the Word Error Rate (WER) on the MyST testset 13.93% to 9.11% with Whisper-Small and from 13.23% to 8.61% with Whisper-Medium and show that this improvement can be generalized to unseen datasets. We also highlight important challenges towards improving children’s ASR performance and the effect of fine-tuning in improving the transcription of disfluent speech.
- Open Access
The Promises and Pitfalls of Using Language Models to Measure Instruction Quality in Education
Assessing instruction quality is a fundamental component of any improvement efforts in the education system. However, traditional manual assessments are expensive, subjective, and heavily dependent on observers' expertise and idiosyncratic factors, preventing teachers from getting timely and frequent feedback. Different from prior research that mostly focuses on low-inference instructional practices on a singular basis, this paper presents the first study that leverages Natural Language Processing (NLP) techniques to assess multiple high-inference instructional practices in two distinct educational settings: in-person K-12 classrooms and simulated performance tasks for pre-service teachers. This is also the first study that applies NLP to measure a teaching practice that is widely acknowledged to be particularly effective for students with special needs. We confront two challenges inherent in NLP-based instructional analysis, including noisy and long input data and highly skewed distributions of human ratings. Our results suggest that pretrained Language Models (PLMs) demonstrate performances comparable to the agreement level of human raters for variables that are more discrete and require lower inference, but their efficacy diminishes with more complex teaching practices. Interestingly, using only teachers' utterances as input yields strong results for student-centered variables, alleviating common concerns over the difficulty of collecting and transcribing high-quality student speech data in in-person teaching settings. Our findings highlight both the potential and the limitations of current NLP techniques in the education domain, opening avenues for further exploration.
Open AccessThe Promises and Pitfalls of Using Language Models to Measure Instruction Quality in Education
Assessing instruction quality is a fundamental component of any improvement efforts in the education system. However, traditional manual assessments are expensive, subjective, and heavily dependent on observers' expertise and idiosyncratic factors, preventing teachers from getting timely and frequent feedback. Different from prior research that mostly focuses on low-inference instructional practices on a singular basis, this paper presents the first study that leverages Natural Language Processing (NLP) techniques to assess multiple high-inference instructional practices in two distinct educational settings: in-person K-12 classrooms and simulated performance tasks for pre-service teachers. This is also the first study that applies NLP to measure a teaching practice that is widely acknowledged to be particularly effective for students with special needs. We confront two challenges inherent in NLP-based instructional analysis, including noisy and long input data and highly skewed distributions of human ratings. Our results suggest that pretrained Language Models (PLMs) demonstrate performances comparable to the agreement level of human raters for variables that are more discrete and require lower inference, but their efficacy diminishes with more complex teaching practices. Interestingly, using only teachers' utterances as input yields strong results for student-centered variables, alleviating common concerns over the difficulty of collecting and transcribing high-quality student speech data in in-person teaching settings. Our findings highlight both the potential and the limitations of current NLP techniques in the education domain, opening avenues for further exploration.