trying to replicate the situation you describe, the data is saved when participant quits (for surveys anyway, I did not try to replicate this for experiments but I suppose the principle is the same).
Best, Matia
The best time to plan an experiment is after you've done it.