Suppose that people who watch a lot of violence on tv are also particularly likely to behave aggressively. this relationship would not necessarily indicate that watching violence influences aggressive behavior because:

The fact that people who watch a lot of violent TV are also more likely to be aggressive does not in itself proof a cause-effect relationship. That is, there is no evidence that one situation is a cause and the other one is a consequence (causation). It only proves that both characteristics are present together (correlation).

This could be a case of “spurious correlation,” which implies that two characteristics are often present together because there is a third factor that influences both of them.  

For example, if we assume that this case refers to minors,  it could be that minors who watch a lot of violent TV are neglected by their parents. Additionally, it could be that children who behave aggressively are also neglected by their parents. If that were the case, then being neglected would be a more likely predictor of violent behaviour, as opposed to watching violent TV.  

In order to make sure there is a cause-effect relationship, we would need to perform an experiment where all other factors are controlled.