Our research
Testing how immersive game-based experiences helps students learn, stay engaged and see university as an option.
Why evidence matters
Good intentions, instinct and anecdotal evidence can point us in useful directions - but they are not enough to tell us reliably what works. Rigorous evidence helps us understand which approaches make a real difference, for whom and under what conditions, and turn that knowledge into better programs, policy and outcomes for students.
Parliament of Australia
I would like to see more randomised trials of financial literacy programs. UNSW’s STEP UP program has already used randomised trials to test its educational impact and is now preparing a large trial of its high-school program. If we are teaching students to demand evidence before investing, our education programs should probably meet the same standard. That kind of evidence can tell us which programs change behaviour and which messages get through. It can also show us what works at different stages of life.
Key research areas
Including gamification, incentives, assessment and educational technology.
Including university aspirations, participation and widening access.
Including accreditation, verifiability and lifelong learning.
Including causal mechanisms, human capital formation and policy simulation.
Including large-scale randomised trials, administrative data and policy translation.
Our findings
Female students whose randomly assigned peer leader was also female scored 0.26 standard deviations higher, an effect comparable to reducing class size by 20%.
Across six trials involving nearly 3,000 students, a more visually engaging learning environment improved female students’ grades by 0.16–0.26 standard deviations, eliminating the gender gap, while male students’ grades were unchanged.
Live rankings, smaller leaderboards and achievement badges lifted course marks. Grouping students with peers at a similar level was especially effective for students who were falling behind.
By Week 2, students’ online engagement could predict whether they would pass or fail with around 85% accuracy, giving teachers an earlier opportunity to provide support.
Medical students who used a video game were more engaged, felt more confident in clinical scenarios (96% compared with 74–76%) and achieved higher learning scores than those using standard materials.
Featured publications
Relative performance feedback in education: Evidence from a randomised controlled trial
Authored by : Loretti Isabella Dobrescu, Marco Faravelli, Rigissa Megalokonomou and Alberto Motta
The power of knowing a woman is in charge: Lessons from a randomised experiment
Authored by : Loretti Isabella Dobrescu, Alberto Motta and Akshay Shanker
Are serious games seriously good at preparing students for clinical practice? A randomised controlled trial
Authored by : Michael Coffey, Loretti Isabella Dobrescu, Alberto Motta, Keith Ooi, et al
Collaborate with us
STEP UP welcomes researchers, educators, policymakers and industry partners to develop and test new approaches to education.
Work with us on experiments, surveys and real-world school programs that explore how to improve educational achievement, student engagement and learning and academic aspirations and what are the channels through which meaningful, lasting change occurs.
Learn more about the data behind our research