A planned decision tree asks: 'Was the target response correct?' then 'Was every correct response reinforced?' Which endpoint should be labeled continuous reinforcement?