Only experiments can provide evidence of a cause-effect link, since in this kind of studies the variables can be controlled in a very specific way.
Hence, the answer is B.
9.5m questions
12.2m answers