The standard test for causality (at least in biology, where I work) is to test for rescue. You first establish that under conditions a, event b happens. You then reverse a, and observe b returning to normal. This, followed by controls demonstrating that you're only changing a and are actually measuring b serve as a stringent test for causation.