The goal of an observational study can be to describe some group, or to examine relationships between variable. 

To deduce cause-and-effect relationship, we must actually do well-designed experiment. 

It is impossible to conclude that one factor causes something to happen in observational study. 

For an example, people think that the level of education and income are correlated. In general, individuals with more years of education are also likely to earn higher incomes, but this may be inaccurate. 

By an experiment in which experimental units are assigned randomly to treatments, we can prove causation.