I totally agree. In fact the systems studied by social sciences are so complex that the notion of an experiment in the classical sense (hypothesis, observations, repeatability, etc) can break down completely. This is partly why the concept of qualitative research exists [1]. Some physical scientists look down on such research methods; this is a category mistake on their part.
[1] https://en.wikipedia.org/wiki/Qualitative_research (not a great article IMO but I lack the experience to give a better reference)