I vaguely remember that the success of "Non-Linear Systems" experiment was initially attributed to people working better when they have "a sense of belonging to a team". However, later it was attributed (by experiment authors, no less) instead to people working better when observed.
