>>12905170Hard science is the domain where you can use the experimental method. That means you have to be able to do controlled, replicable experiments which manipulate only one variable. In a field like chemistry, for example, you can construct precise and repeatable experiments where the only thing you change is the type or amount of a substance
Soft science is the domain where you can use weaker forms of empirical study like correlational studies. In a field like climatology you cannot create a copy of the Earth, make changes, and then compare the effects on the climate compared to the control Earth. The best you can do is collect a bunch of data about changes in the climate over time and then attempt to infer relationships from correlation. In the social sciences you cannot create a perfect copy of a person or a society. You again can only collect a bunch of real world data and try to infer causal effects from the correlations - the best kind of data in psychology is usually a twin study because it gets you closest to a real controlled experiment
Anything else isn't science, although it could still be a useful process or body of knowledge
