The presence of an association does not necessarily imply a causal relationship.
Observing a simple association between two variables - for example, having received a particular treatment, and having experienced a particular outcome - cannot be assumed to mean that the treatment was necessarily responsible for the outcome. The association may reflect the effects of biases from confounder.
Image from XKCD.
For example, people who seek and receive a treatment may be healthier and have better living conditions than people who do not seek and receive a treatment, so the former have better outcomes for those reasons, rather than because of the treatment.
If you feel that this definition hasn't helped you to understand the term, click on our monkey to let us know.