In this case one variable is access to textbooks and the other is exam scores.
However, correlation may not imply causation, especially if there are inherent biases in the samples.