This has been the case of many psychology experiments. There are three possible scenarios:
1. Listening to certain music is strongly correlated with drug use.
2. Doing drugs is strongly correlated to listening to certain music.
3. Some other external factor causes both to be correlated.
You cannot truly imply causation in any case based solely on statistics (correlation does not imply causation). Causation can only be implied with experimentation.
