Many of today’s biggest challenges — averting climate tipping points, reaching herd immunity, mobilising protest — are threshold collective action problems, but principles of behaviour in such settings are understudied. In this talk, I present ongoing research asking why people cooperate in such problems. We run a lab experiment (N = 624) on variants of the public goods game (PGG) compared to a linear PGG baseline, varying the minimum contribution threshold and whether missing it triggers a loss. To explain the resulting behaviour we use the CRISP framework, which treats cooperation as driven by personal norms, (perceived) extrinsic incentives, social expectations, and social preferences. Consistent with previous research, we find that thresholds raise contributions, and losses raise them further. Strikingly, while norms and perceived incentives vary strongly across treatments, their strength in predicting behaviour is similar throughout, pointing to common principles of cooperation across these different settings.