This episode is real people designed, and AI narrated.
Here is a question that should alert every educator today:
if the way we teach programming at university level produces a 68% dropout rate, why are we introducing the same methods to children?
That number is real. In the Netherlands, only 32% of HBO informatica students complete their degree on time. At some institutions, non-completion reaches 84%. The average student who drops out does so after 22 months, not in the first weeks, but after investing nearly two years. Meanwhile, enrollment in ICT programs dropped 7.5% this year alone, the steepest decline across all sectors.
This is not just a university problem. It is a pipeline problem. And the pipeline starts in your classroom.
The research is extensive, and the causes are not what many would assume. The top drivers are:
Today, only 7% of HBO informatica graduates are women, revealing how deeply the belonging gap runs.
You might expect us to say: "the problem is that students sit too much." We did investigate this for the past 3 years. CS students are measurably the most sedentary of all university students, with over half sitting 8+ hours daily. The physiological effects are real: reduced blood flow to the brain, impaired executive function, musculoskeletal strain. And there is a documented link between prolonged sitting and worsened mental health.
But here is where intellectual honesty matters: the direct connection between sedentary learning and academic dropout has not been proven. Roughly half of well-designed studies find no meaningful link between physical activity and academic outcomes. The evidence is suggestive, not conclusive.
What the latest research proves beyond doubt is that how programming is taught changes everything. When universities replaced passive lectures with active, collaborative methods, failure rates dropped by up to 55%. Pair programming raised course continuation from 67% to 85%. Problem-based learning cut dropout from 45% to 17%. Small classes (under 30 students) achieve 80% pass rates versus 65% in larger ones.
Beyond the Netherlands, the pattern is consistent across 15 countries and hundreds of studies: students who learn by doing, together, in meaningful contexts, stay.
Now consider: if autonomous AI can increasingly write code, the argument for teaching children to type syntax grows weaker by the day. Software development job postings have fallen 71% since 2022. But the argument for teaching computational thinking (problem decomposition, pattern recognition, logical reasoning) grows stronger. These are the skills AI cannot replace, and they are the skills that active, embodied pedagogy develops best.
You already know the answer. Curriculum pressure, limited resources, the assumption that "tech education" means screens and keyboards. But the evidence now says otherwise. The most effective programming education is collaborative, physically engaging, and focused on thinking, not typing.
As educators, you do not need to be Python experts to facilitate computational thinking learning. What your students need is exactly what good teachers already do: engage curiosity, facilitate collaboration, and create environments where thinking is active, not passive.
