The perfectly replicable study