The University of Memphis made this announcement earlier today:
MEMPHIS, Tenn. — The University of Memphis has partnered with the Educational Testing Service (ETS) Research Institute and Georgia State University on a new $4 million U.S. Department of Education (ED) grant. The team’s research will look at reinventing how college students learn and demonstrate knowledge. The award, announced by ED’s Fund for the Improvement of Postsecondary Education (FIPSE), will help develop AI-enhanced assessments designed to move beyond traditional testing models.
“I’ve been working on innovating assessments with my colleagues for decades,” said Dr. John Sabatini, distinguished research professor at UofM. “Our insights from the learning sciences, however, have mostly been aimed at K-12 education. This grant has given us an opportunity to work together on a new challenge, higher education.”
The project Sabatini is helping to lead builds directly on a 2024 award from the Institute of Education Sciences (IES) that was awarded to the UofM, who at the time was already working alongside ETS and Georgia State. The IES Transformative Research grant — a highly competitive program that funded only three projects nationwide in its most recent cycle — laid the foundation for the team’s rapid success in securing this new $4 million investment.
The team’s work asks a simple but profound question: Are students learning, or are they just performing well enough on multiple-choice, fill in the blank and essay exams to move on? These scenario-based assessments (SBAs) aim to close that gap by placing students in learning environments that mirror how people acquire and use knowledge in the real world.
“The concept of performance assessment, a precursor of SBAs, has been around for many years,” said Sabatini. “But in practice, they were too costly to develop, time-consuming to implement and score and often lacked measurement integrity. With the widespread availability of digital technologies — and now AI — the landscape has changed. What was impractical or infeasible then is fast becoming a necessity of quality instruction today.
Sabatini, who joined UofM in 2019, has been at the center of this research. A former senior researcher at ETS with 17 years of experience, he’s helped shape a research agenda focused on SBAs — a fundamentally different way of understanding whether students truly learn, retain and apply knowledge.
The IES Transformative grant project, now in its second year, will design and test 12 SBAs, using data from students who enroll in three of the most widely taken courses in American higher education: Psychology 101, English Composition 101 and Interdisciplinary Studies. These courses enroll millions of students each year, making them ideal proving grounds for a new assessment framework that could potentially reshape how colleges measure learning.
The SBAs, according to Sabatini, will incorporate multimodal elements, which will include text, images, video and oral responses, to better reflect how students learn outside the classroom. The goal is to help students build durable knowledge they can carry into their personal, professional and graduate-level lives.
“We are on the cusp of a multimodal media revolution,” said Sabatini. “Multimodal literacy has emerged as a means for storing and communicating knowledge that is potentially as powerful and universal as the printed word has been for the past 500 years. We can use digital technologies to access and respond in the four modalities of language: speaking, listening, reading via text-to-speech, writing via speech-to-text in real time or remotely. We can produce visual and video content on demand. We now need to integrate these modalities into learning and, as importantly, into the ways we assess students.”
He notes that the timing of this work is critical. With generative AI reshaping nearly every corner of education, the research team sees a rare opportunity to rethink assessment from the ground up.
Despite the project’s technological ambitions, Sabatini emphasizes that the team is not composed of computer scientists. Instead, it brings together psychologists, literacy experts, learning scientists and educators — a multidisciplinary group focused on how humans learn. That perspective, he says, is essential to ensure that AI tools enhance learning rather than simply automate old models.
“There is unfortunately a trend toward using AI to become more efficient at generating multiple-choice questions and scoring on-demand writing,” Sabatini said. “AI is great at supporting both. While we welcome efficiencies, we worry that exciting new and authentic ways to learn will remain secondary to traditional instruction, especially if the testing remains the same. That’s why we want to change assessments as a means toward empowering innovative learning.”
Joining Sabatini from the UofM are Drs. Randy Floyd, Mollie Anderson and Emily Srisarajivakul from the Department of Psychology; Drs. Katherine Fredlund and J. Elliot Casal from the Department of English; and Dr. Karen Weddle-West from the College of Education.
Author
