A format I developed for subnet: data literacy as play. Teams take real datasets and set out to "prove" the most absurd hypothesis they can — cherry-picking, p-hacking, axes bent just so — and then take each other's posters apart. To see how selective data and misleading visualization actually work, you build them yourself.

We live in a world that argues with models and statistics, yet hardly anyone runs a genuine data-based discourse. Two toxic extremes rule instead: blind faith in numbers, and fear of maths. The Jam goes at both — it deliberately perverts the scientific process so you can see through it, in a protected frame, with mentors calling each move by its name: "that's p-hacking", "that's survivorship bias".

It is the playful flip-side of plankton. Where plankton makes a claim's derivation checkable, the Jam makes the tricks of manipulation tangible by having you commit them yourself — fact-checks answer single lies; the Jam inoculates against the method.

I developed it for subnet; the first iteration is a prototype run in cooperation with Schmiede Hallein (16–26 September 2026), with methodical anchoring by the Plankton Research Group. Erik Zepka joins as a second methodologist, and criminologist Walter Fuchs brings quantitative social-science rigour. The format is still in development — this is an orientation, not a finished thing, and it has not happened yet.

To orient:

The full format, day by day — planktonresearch.org → At the Schmiede — schmiedehallein.com →