- Press Release Distribution

Study: Algorithms used by universities to predict student success may be racially biased

Predictive algorithms underestimate the likely success of Black and Hispanic students

( Washington, July 11, 2024—Predictive algorithms commonly used by colleges and universities to determine whether students will be successful may be racially biased against Black and Hispanic students, according to new research published today in AERA Open, a peer-reviewed journal of the American Educational Research Association. The study—conducted by Denisa Gándara (University of Texas at Austin), Hadis Anahideh (University of Illinois Chicago), Matthew Ison (Northern Illinois University), and Lorenzo Picchiarini (University of Illinois Chicago)—found that predictive models also tend to overestimate the potential success of White and Asian students. 

Video: Co-authors Denisa Gándara and Hadis Anahideh discuss findings and implications of the study

“Our results show that predictive models yield less accurate results for Black and Hispanic students, systemically making more errors,” said study co-author Denisa Gándara, an assistant professor in the College of Education at the University of Texas at Austin.

These models incorrectly predict failure for Black and Hispanic students 19 percent and 21 percent of the time, respectively, compared to false negative rates for White and Asian groups of 12 percent and 6 percent. At the same time, the models incorrectly predict success for White and Asian students 65 percent and 73 percent of the time, respectively, compared to false negative rates for Black and Hispanic students of 33 percent and 28 percent.

“Our findings reveal a troubling pattern—models that incorporate commonly used features to predict success for college students end up forecasting worse outcomes for racially minoritized groups and are often inaccurate,” said co-author Hadis Anahideh, an assistant professor of industrial engineering at the University of Illinois Chicago. “This underscores the necessity of addressing inherent biases in predictive analytics in education settings.”

The study used nationally representative data spanning 10 years from the U.S. Department of Education’s National Center for Education Statistics, including 15,244 students.

Findings from the study also point to the potential value of using statistical techniques to mitigate bias, although there are still limitations.

“While our research tested various bias-mitigation techniques, we found that no single approach fully eliminates disparities in prediction outcomes or accuracy across different fairness notions,” said Anahideh.  

Higher education institutions are increasingly turning to machine learning and artificial intelligence algorithms that predict student success to inform various decisions, including those related to admissions, budgeting, and student-success interventions. In recent years, there have been concerns raised that these predictive models may perpetuate social disparities.

“As colleges and universities become more data-informed, it is imperative that predictive models are designed with attention to their biases and potential consequences,” said Gándara. “It is critical for institutional users to be aware of the historical discrimination reflected in the data and to not penalize groups that have been subjected to racialized social disadvantages.”

The study’s authors noted that the practical implications of the findings are significant but depend on how the predicted outcomes are used. If models are used to make college admissions decisions, admission may be denied to racially minoritized students if the models show that previous students of the same racial categories had lower success. Higher education observers have also warned that predictions could lead to educational tracking, encouraging Black and Hispanic students to pursue courses or majors that are perceived as less challenging.

On the other hand, biased models may lead to greater support for disadvantaged students. By falsely predicting failure for racially minoritized students who succeed, the model may direct greater resources to those students. Even then, Gándara noted, practitioners must be careful not to produce deficit narratives about minoritized students, treating them as though they had a lower probability of success.

“Our findings point to the importance of institutions training end users on the potential for algorithmic bias,” said Gándara. “Awareness can help users contextualize predictions for individual students and make more informed decisions.”

She noted that policymakers might consider policies to monitor or evaluate the use of predictive analytics, including their design, bias in predicted outcomes, and applications.

Funding note: This research was supported by the Institute of Education Sciences at the U.S. Department of Education

Study citation: Gándara, D., Anahideh, H., Ison, M., & Picchiarini, L. (2024). Inside the black box: Detecting and mitigating algorithmic bias across racialized groups in college student-success prediction. AERA Open, 10(1), 1–15.


About AERA
The American Educational Research Association (AERA) is the largest national interdisciplinary research association devoted to the scientific study of education and learning. Founded in 1916, AERA advances knowledge about education, encourages scholarly inquiry related to education, and promotes the use of research to improve education and serve the public good. Find AERA on Facebook, X, LinkedIn, Instagram, Threads, and Bluesky.



Comprehensive evaluation of large language models in mining gene relations and pathway knowledge

Comprehensive evaluation of large language models in mining gene relations and pathway knowledge
Understanding complex biological pathways, such as gene-gene interactions and gene regulatory networks, is crucial for exploring disease mechanisms and advancing drug development. However, manual literature curation of these pathways cannot keep pace with the exponential growth of discoveries. Large-scale language models (LLMs) trained on extensive text corpora contain rich biological information and can be leveraged as a biological knowledge graph for pathway curation. Recently, Quantitative Biology published a study titled "A Comprehensive ...

Researchers pinpoint brain cells that delay first bite of food

Researchers pinpoint brain cells that delay first bite of food
LA JOLLA, CA—Do you grab a fork and take a first bite of cake, or say no and walk away? Our motivation to eat is driven by a complex web of cells in the brain that use signals from within the body, as well as sensory information about the food in front of us, to determine our behaviors. Now, Scripps Research scientists have identified a group of neurons in a small and understudied region of the brain—the parasubthalamic nucleus (PSTN)—that controls when an animal decides to take a first bite of food. In the study, published in Molecular Psychiatry on July 4, 2024, the team of scientists set out to selectively manipulate a group of PSTN cells that dial up their ...

With spin centers, quantum computing takes a step forward

With spin centers, quantum computing takes a step forward
RIVERSIDE, Calif. -- Quantum computing, which uses the laws of quantum mechanics, can solve pressing problems in a broad range of fields, from medicine to machine learning, that are too complex for classical computers. Quantum simulators are devices made of interacting quantum units that can be programmed to simulate complex models of the physical world. Scientists can then obtain information about these models, and, by extension, about the real world, by varying the interactions in a controlled way and measuring the resulting behavior of the quantum simulators.   In a paper published in Physical Review B, a UC Riverside-led research team ...

Scientists release new research on planted mangroves’ ability to store carbon

Scientists release new research on planted mangroves’ ability to store carbon
U.S. Forest Service ecologists and partners published new findings on how planted mangroves can store up to 70% of carbon stock to that found in intact stands after only 20 years. Researchers have long known that mangroves are superstars of carbon absorption and storage. But until now, limited information existed on how long it took for carbon stored in planted mangroves to reach levels found in intact mangroves. “About ten years ago, Sahadev Sharma, then with the Institute of Pacific Islands Forestry, and I discovered that 20-year-old mangrove plantations in Cambodia had carbon stocks comparable to those of intact forests,” ...

New immune cell therapy benefits laboratory models of ALS and has some positive results in an individual with the disease

Immune system dysregulation and elevated inflammation contribute to the development of the fatal neurodegenerative condition amyotrophic lateral sclerosis (ALS), also known as Lou Gehrig's disease.  In new research published in The FASEB Journal, repeated infusions of certain immune cells delayed ALS onset and extended survival in mice, and also reduced markers of inflammation in an individual with the disease. The work was conducted by investigators at Massachusetts General Hospital, a founding member of the Mass General ...

Trial of cell-based therapy for high-risk lymphoma leads to FDA breakthrough designation

CAR-T cell therapy, which targets a specific protein on the surface of cancer cells, causes tumors to shrink or disappear in about half of patients with large B-cell lymphoma who haven’t experienced improvement with chemotherapy treatments. But if this CAR-T treatment fails, or the cancer returns yet again — as happens in approximately half of people — the prognosis is dire. The median survival time after relapse is about six months. Now, a phase 1 clinical trial at Stanford Medicine ...

Major trial looks at most effective speech therapy for people with Parkinson’s disease

A major clinical trial, led by experts at the University of Nottingham, has shown the Lee Silverman Voice Treatment (LSVT LOUD®) is more effective than the current speech and language therapy provided by the NHS, when treating patients with Parkinson’s disease (PD). The results of the NIHR HTA funded trial, which are published today in the BMJ, showed that LSVT LOUD® was more effective at reducing the participant’s reported impact of voice problems than no speech and language therapy, as well as the NHS delivered speech and language therapy. The trial was led by experts from the Universities of Nottingham and Birmingham, ...

Intensive voice treatment more effective than NHS speech therapy for Parkinson’s disease

An intensive voice treatment developed in the USA and known as the Lee Silverman voice treatment (LSVT LOUD) is more effective than conventional NHS speech and language therapy or no therapy for people with Parkinson’s disease, finds a trial published in The BMJ today. The researchers say the results emphasise the need to optimise the use of speech and language therapy resources for people with Parkinson’s disease. Slurred or slow speech (known as dysarthria) is a common feature of Parkinson’s disease and can have a significant effect ...

NHS’s relentless focus on finance and productivity is failing patient safety

“The NHS’s relentless focus on finance and productivity is failing patient safety,” argues patient safety commissioner Henrietta Hughes in an interview for The BMJ today. “The patient’s anecdote is the canary in the coal mine,” she says. It’s the thing that tells us there’s something going wrong. But too often we hear about patients who have raised concerns being gaslighted, dismissed, and fobbed off.” Hughes, a GP and former medical director at NHS England, became England’s first independent patient safety commissioner in September 2022. The job was created ...

World leaders unite to embed social participation in health systems

For the first time in the World Health Organization’s 76 year history, world leaders have unanimously committed to put social participation - people, communities and civil society - at the heart of health decision making processes. This landmark pledge marks an important step forward in creating a healthy world, say experts in The BMJ today, and they urge everyone to seize their right to influence the decisions that affect their health and well-being. Global challenges such as access to care, climate change, widening ...


Duke-NUS launches LIVE Ventures, a S$20 million incubator to accelerate research commercialisation

Samuel Pepys’ fashion prints reveal his guilty pleasure: Fancy French clothes

New genetic test will eliminate a form of inherited blindness in dogs

Cancer risk: Most Australian welders exposed to high levels of dangerous fumes

Two-in-one mapping of temperature and flow around microscale convective flows

Texas A&M engineers explore intelligence augmentation to improve safety

ORNL economist honored at international hydropower conference

UCLA selected by Centers for Medicare & Medicaid Services to test Medicare dementia care model

Fish adjust reproduction in response to predators

DDX41 and its unique contribution to myeloid leukemogenesis

Digital games on vaping devices could lure more youth to nicotine addiction

Cracking the code of hydrogen embrittlement

Long-term results from Testicular Cancer treatment are positive, study shows

EPA awards UMass Amherst nearly $6.4 million to help shrink the steel industry’s carbon footprint

Valentina Greco takes on new position as President of the ISSCR

Komen supports UVA Engineering researchers targeting ‘triple negative' breast cancer

Panel issues first guidelines to prevent anal cancer in people with HIV

Estimating rainfall intensity using surveillance audio and deep-learning

Targeting factors for chemoprevention and cancer interception to tackle mesothelioma

New snake discovery rewrites history, points to North America’s role in snake evolution

Large and unequal life expectancy declines in India during COVID-19

A study of 156,000 UK residents found that urban residents score the lowest in social and economic satisfaction and well-being

Global study by Hawaiʻi Institute of Marine Biology demonstrates benefit of marine protected areas to recreational fisheries

Researchers clarify how soft materials fail under stress

Revolutionizing the abilities of adaptive radar with AI

Plastic waste can now be converted to electronic devices

Health equity scholar Darrell Hudson named Health Behavior and Health Education chair at the University of Michigan School of Public Health

Research will establish best ‘managed retreat’ practices for communities faced with climate change disaster

Marshall University awarded grant to further fentanyl addiction research

Wash U researchers shine light on amyloid architecture

[] Study: Algorithms used by universities to predict student success may be racially biased
Predictive algorithms underestimate the likely success of Black and Hispanic students