Re:Address to the American Psychological Association on Men ...
by Ended on Friday November 11, 2016 @05:29AM
(#12929455)
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)
To a large extent progress in sciences like sub-atomic physics depends on ability to (1) simplify the problem down to mathematical laws, (2) conduct repeatable, randomized controlled experiments to rule out hypotheses about them.
Social sciences deal with phenomena that are extraordinarily complex, so hypotheses usually have to be qualitative rather than mathematical, and devising good experiments is very hard. Falsification is hard. Most hypotheses can be "generally true" yet have specific counterexamples in some times and places.
As for empirical data, you can use data to tell a million plausible stories, and many of those probably have some nugget of truth. But without randomized trials, how do you falsify any of these stories? It's hard.