Yes, I do think social science is largely bullshit 
especially when it tries to use numbers to claim something.
That said with respect to gun studies in particular - here's why it isn't falsifiable and thus bull plop . Suppose I'm charged with investigating a small number of people who claimed to have had a defensive gun use in the last year. Further suppose that the people I'm charged with interviewing are named Tom, Dick, and Harry.
Tom did have a legitimate defensive gun use last year. He was minding his own business on a walk and was assaulted by 3 drunk morons and scared them off by pulling his gun.
Dick didn't have a defensive gun use. He lied because he thinks people are coming to take his guns.
Harry did pull his gun on someone, but it was another motorist, after he cut them off and they flipped him off and honked at him, but otherwise indicated no violent intent.
So I interview them. Tom has a good truthful story.
Dick claims someone tried to carjack him up at a gas station and he scared them off.
Harry claims a crazy motorist almost ran him off the road several times until he pulled his gun.
There is absolutely no way I could prove Dick or Harry are lying and/or mistaken about having a defensive gun use unless a GoPro had been on their heads at all times and I could see the incident in question.
Edit history
= new reply since forum marked as read