[kergonath]

> it is obvious social studies can not have the rigor of Newtonian physics simply by the difference in running experiments for each discipline.

To my mind (trained in statistical Physics), you don’t have to abandon rigour when looking at complex systems. We can build models for things we don’t understand perfectly and we can rationalise experiments even if we don’t control everything (we rarely ever do). I think it’s a difference in degree rather than nature.

The fact that many sociological studies are not reproducible does not mean that we cannot learn anything by observing and investigating correlations. It means that we need to be more serious while doing it.

> You really can't take a thousand babies, isolate them and run experiments to see what is a social construct and what is not.

You’re right. But this means that the emphasis moves from the experimental setup to the data analysis, and also that the results need to be discussed in terms of trends and likelihood more than verified facts. The field could still do away with p-hacking, data fabrication, and dodgy ideologies.

There are plenty of physicists who publish rubbish because they set out to prove a pet theory and become victims of confirmation bias. That tendency is present everywhere, it is just stronger in fields where you cannot control as much the system you observe.