It is very difficult to do scientific research with people. People often do things that defy explanation. Although we can do research with people, and we can make generalizations, it's difficult for us to create laws to describe their behavior because people can do unexpected things. When doing research involving statistics, you also have to be a bit skeptical because it's easy to manipulate numbers.
One reason that this is difficult is that situations that social scientists try to study have too many variables. Because there are so many variables, it is hard to isolate the impact of just one of those variables the way that scientists can while working in a laboratory. In addition those variables can be difficult to measure. An example of this could be studies of crime.
For example, let's imagine you wanted to study the impact of the quality of education on crime rates. You hypothesize that an area with bad schools would have high crime rates. You would find at least two problems. First, you would have a hard time defining and measuring the quality of education in a given school or school system. Second, there are many other things that might have an impact on crime. For example, what if unemployment rates went up right as education got worse? Which thing would you blame an increase in crime on in that circumstance?
It is for reasons like this that precise laws about social sciences cannot be formulated. The variables are hard to measure and there are too many of them to allow really precise laws to be discovered.