open-science-lab / cambridge

Open Science Lab

I built these tools for my teaching across undergraduate and MPhil courses at the University of Cambridge. They started as in-lecture activities for supervisions and lab sessions, and I'm sharing them here so others can use them.

Activities

Multiple Comparisons Simulation

Run 5, 20, or 100 statistical tests on data where nothing is real. Watch false positives accumulate and see why correction methods like Bonferroni matter.

simulation false positives multiple testing

Spot the Bad Practice

Five research scenarios drawn from neuroscience and psychology. Identify whether each one is p-hacking, HARKing, publication bias, or data fabrication.

quiz research integrity QRPs

Garden of Forking Paths

Same dataset, different analysis choices. Make five defensible decisions about outliers, covariates, and statistical tests — then see how many of the 32 possible paths produce a "significant" result.

exploration researcher degrees of freedom pre-registration

About this project

These activities grew out of a recurring problem: teaching the replication crisis and questionable research practices through slides alone doesn't land with students the way actually experiencing them does.

I use these in lectures, supervisions, and small-group teaching for PBS03 and related methods courses. Everything runs in the browser with no setup required — students just open the link on their phone or laptop.

All materials are open source under the MIT licence. If you teach research methods and find these useful, feel free to fork, adapt, or get in touch.