I've seen a sociologist's tool for "textual narrative analysis" at work - the thing that they use to compare data from transcribed interviews with people. It's a pretty hardcore little piece of software, and is not unlike its new big brothers, the big data analytic engines.
Using technology doesn't make what you're doing science.
Trust me, they're being as scientific about that kind of stuff as any other scientist out there. They use as much statistical analysis as any botanist or chemist.
Not even close to the same amount. Statistics and math are almost afterthoughts in sociology course work.
They use the scientific method - propose a hypothesis, conduct experiments (interviews usually), compile the results, compare the results to the hypothesis, and draw new conclusions or re-affirm old conclusions as necessary.
On rare occasions, I'm sure this does happen. Most of the time it's testing to the theory (and I'm using "theory" loosely here), or deriving the theory post-hoc from the testing (with follow up testing rarely occurring). To be good science one has to derive a test to disprove the hypothesis. Sociological experiments very very very rarely do this.
They have strict controls over what is acceptable usage of data, who they can interview, etc done by the IRB boards for their respective universities and think tanks. (They can't interview minors without parental approval, for example, unless explicitly permitted.)
Which, again, has nothing to do with whether they're using proper scientific methodology.
The only difference is that their test subjects are human beings instead of plants or chemicals.
Oh, I fully agree with the increased difficulties involved, but the reality is that most people gravitate into a sociology track in college precisely because they want a field that's less rigid with the methodology and math and the other things that make other subjects "hard". The result is a marked difference, and one of the reasons why the field is largely considered a joke among science.