Causal Inference in Educational Data Mining
EDM 2024 Half-Day Workshop
Workshop Overview
The goal of crafting intelligent tutoring systems, educational games, MOOCs, and similar computerized learning tools is to enhance student learning. While the primary focus of Educational Data Mining (EDM) research remains on techniques for identifying, measuring, and predicting learner behaviors or outcomes, there is a critical need to understand causality. Causal inference goes beyond mere prediction to estimate the impacts of various factors on students' behaviors or learning outcomes. It is not just about predicting who will struggle or succeed, but unraveling the underlying causes behind these phenomena. Causality plays a pivotal role both in learning science, which explores how educational interventions and computerized inputs affect educational outputs, and in policy-making, where the aim is to design and systematically deploy systems that enhance learning. Causal inference allows us to confidently make design decisions for educational technology to improve learning outcomes.
The domain of causal inference, encompassing fields of statistics, philosophy, economics, and computer science, has seen rapid advancements. This emerging science addresses the challenges involved in estimating effects amidst complex situations in which confounding variables can obscure results. Thus, it offers methods that allow for examining causality within the context of digital learning platforms which further affords opportunities that may leverage the utilization of educational data mining. Using causal methods, we can explore how impacts of learning interventions and educational technologies differ among learners and identify the mechanisms driving causal effects. Additionally, there is great potential for utilizing educational data mining methodologies to strengthen causal inference techniques within educational contexts.
In sum, the workshop will be organized to stimulate discussion among participants, including, hopefully, constructive suggestions for open problems related to causal inference in the context of EDM research.
Call for Papers/Presentations
Our workshop aims to shed light on the prevalence and significance of causal inquiries within EDM while emphasizing the two overarching themes of examining causality within EDM contexts and leveraging EDM to support causal inference methodologies.
We welcome completed, ongoing, and early-stage work on topics including, but not limited to:
A/B testing
Graphical causal models/Bayesian networks
Analyzing data from randomized experiments
Multi-armed bandits
Investigations of causal mechanism/mediation analysis
Estimating EDM program impacts
Identifying and predicting differential effects
Connections between machine learning and causal inference
Dynamic treatment regimes
Principal stratification
Causal inference in EDM without randomization
Submission Types
Full Papers: up to 10 pages | 15 minute presentation
Short Papers: up to 6 pages | 10 minute presentation
Open Problems/Opportunities for EDM*: up to 4 pages
*Presentation slides for Open Problems/Opportunity papers are optional but presentors will lead discussion
Reviewing and Proceedings
The workshop organizers will review submissions, alongside ad hoc external reviewers whose expertise is appropriate for the submissions. Accepted submissions will be shared with participants prior to the workshop and made available through this web page following (any author/presentor may request for their paper to not be shared publicly following the workshop).
Submission Guidelines
Papers should follow the EDM conference format and submissions are accepted through easychair.
Submissions are due by June 10, 2024 11:59pm AOE
Tentative Schedule*
*The workshop schedule will be updated based on the number of accepted papers for each submission type
0:00-0:10 -- Introduction
0:10-1:10 -- Keynote Speaker (TBA)
1:10-2:00 -- Long Paper Presentations (15 minutes each with 10 minute discussions)
2:00-2:45 -- Short Paper Presentations (10 minutes each with 5 minute discussions)
2:45-3:30 -- Session on Open Problems and Opportunities for EDM
Anthony F. Botelho
Anthony is an Assistant Professor of Educational Technology and computer science education in the College of Education at the University of Florida. His research seeks to impact learning through the blending of learning theory and quantitative methods. Anthony's primary lines of research include the study of student cognition, behavior, and affect, identifying effective learning interventions through causal inference, and developing human-in-the-loop systems and tools to support teachers.
Avery H. Closser
Avery is an incoming Assistant Professor of Emerging Technologies and Learning in the College of Education at the University of Florida. Her research aims to leverage cognitive theory to advance learning technologies and open materials for instructional practice. She specializes in experimental design in the context of learning technologies and explores best practices for methodologies related to this area of research.
Adam C. Sales
Adam is an Assistant Professor of Mathematical Sciences and an affiliate of the Learning Sciences and Technologies and Data Science programs at WPI. His research in applied statistics focuses on methods for causal inference using large, administrative datasets, primarily with applications in learning sciences and social sciences. He has developed and worked on methods combining machine learning with design-based analysis of randomized trials and matched observational studies, principal stratification and mediation analysis using log data from intelligent tutoring systems, and regression discontinuity designs.
Neil T. Heffernan
Neil is the William Smith Dean's Professor of Computer Science at WPI, the creator of ASSISTments, and an active researcher in the fields of 1) artificial intelligence and education, 2) educational data mining and 3) learning analytics. In order to support research in these fields, Dr. Heffernan created the E-TRIALS Testbed, a tool that allows ASSISTments to be used as a platform to do science and support evidence-based practice. He has dozens of papers in educational data mining, and 20+ papers in comparing different ways to optimize student learning.
Kirk P. Vanacore
Kirk is a Ph.D. student in Learning Sciences and Technologies. He applies statistical and machine learning models to data from computer-based learning platforms to understand learning mechanisms. His work includes using experimental designs and observational studies to understand the nuances of how educational programs and pedagogies impact learning. He focuses on the intersection of failure, struggle, and learning.