Causal relationships
In many bivariate data sets, the relationship between the two variables is not symmetric. From the nature of the variables and the way that the data were collected, it may be clear that one variable, X, can potentially influence the other, Y, but that the opposite is impossible.
In such data, the variable X is called the explanatory variable and Y is called the response.
Experiments
In an experiment, the person conducting the experiment controls the values of the explanatory variable. A well-designed experiment always ensures that the relationship between the explanatory variable and response is causal.
Observational studies
If the person collecting the data has no control over either of the variables, and simply records a pair of values from each individual, then the data are called observational. If one variable is an earlier measurement than the other, we may also be able to treat it as an explanatory variable and the later variable as the response.
Even if the relationship is not causal, we are sometimes interested in predicting the value of one variable from the other. The variable being predicted would then be treated as the response.