The reason you have psychology experiments with controls and parameters is that extracting definite conclusions from the simple observation of human behavior is extremely difficult given the wide variety of individuals, groups and cultures.

Once you have an experiment that degenerates into just an event, a situation where the controls have failed, you come up with many potential conclusions but you've lost any science-specific-conclusion to the observations and you may as well look any series of events.

That said, I think experimental psychology just generally fails to establish enough controls to merit the scientific quality it aspires to.