From data we may observe that two phenomena are related, even that one always precedes the other. However, it can sometimes be hard to establish causality - which is the cause of the other, the reason it happens or has the value it does. Sometimes two phenomena may be linked by a common cause rather than one being the cause of the other. For example, from data we might note that people who do a lot of exercise are also more healthy, and so think that exercise causes good health, but equally well it might be that people who are in good health are more likely to take exercise, or that there is some aspect of people's upbringing or genes that leads to both. Understanding mechanism can sometimes help establish causality.
Used on page 34