While research and practice centred around students and academics working together to co-create in the higher level sector has increased, co-creation in assessment remains relatively rare in a higher education context. It is acknowledged in the literature that deeper comprehension of content can be realised when students author their own questions and solutions, rather than just answering teacher-designed questions. However, to date, studies measuring the impact of students co-creating assessment instruments on academic performance are limited. Situated within a constructivist learning paradigm, this study examines the effect on academic performance of students creating their own multiple-choice questions using an on-line peer-learning environment. We use a novel experimental approach, designed to overcome many of the acknowledged limitations of prior studies, to quantitatively measure the effectiveness of the co-creation assignment on student learning. We find that co-creation has a statistically significant impact on their academic performance in relevant topic in the end-of-semester exam.