Brilliant.org was founded in 2012 and spent several years growing a reputation as the most intellectually serious of the self-paced STEM learning platforms: where Khan Academy's videos were accessible and systematic, Brilliant's approach was active, problem-first, and deliberately harder. By 2019, the platform had a substantial community of learners and a curriculum spanning mathematics, physics, computer science, and data analysis, built around the premise that deep understanding came from working through problems, not from passively watching explanations.
Whether that approach served girls and women in STEM better or worse than the alternatives was not a question that Brilliant had answered publicly with rigorous evidence by 2019. Nor had it been answered by independent researchers in a way that clearly favoured one self-paced platform over another on gender grounds. What was available was a set of theoretical arguments, some indirect evidence from related research, and an honest engagement with what the platform's design actually implied.
The Problem-Based Approach and Gender Research
The research literature on gender and mathematics education had found, with some consistency, that girls tended to respond better than boys to pedagogical approaches that emphasised mathematical reasoning, conceptual understanding, and problem-solving over rote procedure application — a finding that, if applied to Brilliant's design, might suggest the platform was well-suited to girls. This was not, however, a finding that straightforwardly endorsed any particular platform design; it was a finding about pedagogical approach at the classroom level, and the translation to a self-paced digital environment raised questions about what "problem-based learning" meant when no teacher was present.
The specific challenge with pure problem-based approaches in self-directed contexts was productive struggle — the point at which a problem was hard enough to require persistence and hard enough that persistence without feedback could tip into frustration and disengagement. Research on girls' engagement in STEM education had found that the relationship between difficulty and disengagement was not straightforward: girls were not uniformly more likely to disengage at difficulty, but they were more sensitive to certain kinds of difficulty framing — difficulty that felt like evidence of not belonging (because the surrounding signals said this domain was not for them) was experienced differently from difficulty that felt like evidence of a challenging problem worth solving.
Brilliant's problem-curation approach — selecting problems at the frontier of what a learner could solve with genuine effort — was designed to make the second kind of difficulty the dominant experience. Whether it succeeded in doing this for girls who had received extensive messages that mathematics and physics were not domains in which they belonged was an empirical question that the platform had not published evidence on.
Comparison with Khan Academy
Khan Academy was, by 2019, the largest self-paced mathematics education platform in the world by user base — its video-explanation model, introduced when Salman Khan began tutoring his cousins remotely in the mid-2000s, had scaled to reach hundreds of millions of learners. The pedagogical contrast with Brilliant was significant: Khan Academy's core offering was worked examples and explanatory videos followed by practice exercises; Brilliant's was the inverse — problems first, with hints and explanations available on demand.
The evidence on gender and Khan Academy's effectiveness was similarly indirect. Khan Academy's massive scale meant that it had accumulated engagement data that it had used internally to improve learning pathways; the public reporting on gender-disaggregated learning outcomes was limited. What the platform had published showed completion rates that varied by subject and by user demographic in ways that reflected broader educational equity patterns, but gender-specific outcome data was not widely available.
The pedagogical question of whether video-first or problem-first approaches served girls differently was not, by 2019, settled. The research literature on video-based mathematics instruction and gender had produced mixed findings. Girls' preferences for different instructional formats varied with prior knowledge, with the subject domain, and with the specific way difficulty was handled in each format. Claiming that either platform clearly served girls better required more direct evidence than was available.
What Self-Paced STEM Platforms Cannot Do
Both Brilliant and Khan Academy, like all self-paced digital learning platforms, operated in a space of individual learner engagement. They could not change the social environment in which a girl learned mathematics — the classroom where she sat, the family expectations about her academic potential, the cultural messaging about who mathematics was for. The research evidence on women's persistence in STEM consistently found that social belonging was a more powerful determinant of persistence than any pedagogical feature of the learning experience.
For WIGSAT's audience — women and girls in STEM globally — the gap between what self-paced platforms could deliver and what structural equity in STEM education required was significant. A teenage girl in Kampala who used Brilliant to learn calculus was developing a skill; she was not thereby changing the conditions at the university she would apply to, the reception she would receive in the engineering faculty if she were admitted, or the career trajectory available to her once she graduated. The platform served the individual dimension of the problem; the structural dimension was not within its scope.
The Global Access Question
Brilliant.org's model required reliable internet access and a device capable of rendering interactive content — constraints that were not trivially met across all of WIGSAT's geographic scope. A self-paced learning platform designed for learners in OECD countries with robust connectivity was a different tool in the hands of a learner in a rural Indian state or sub-Saharan African country with intermittent connectivity and shared devices. The gap between a platform's global availability (accessible via a browser anywhere) and its global accessibility (genuinely usable in low-connectivity contexts with varying device quality) was a genuine limitation for any approach to global STEM equity that relied primarily on digital platforms.
Frequently Asked Questions
What is Brilliant.org?
Brilliant.org is a self-paced STEM learning platform founded in 2012, focused on mathematics, physics, computer science, and data analysis. Its distinguishing approach is problem-first learning — presenting learners with problems to work through rather than explanatory content to watch passively — with a curriculum curated for intellectual rigour.
Does Brilliant.org serve girls' STEM development better than video-based platforms like Khan Academy?
The evidence does not clearly establish that either approach is superior for girls. The research on problem-based learning and gender is theoretically supportive of problem-first approaches, but the translation to self-paced digital contexts introduces confounds that the classroom-level research does not resolve. Neither platform has published rigorous gender-disaggregated outcome data.
What are the limits of self-paced STEM platforms for girls' STEM equity?
Self-paced platforms can develop individual skills but cannot change the social and structural conditions — classroom culture, family expectations, institutional climate — that research identifies as the more powerful determinants of girls' STEM persistence. They also require connectivity and devices that are not equally available globally, limiting their reach in the low-income contexts where STEM access gaps are most acute.
← Girls Who Code Launches | ← Code.org: Five Years | The Pandemic and Women's Research Output →