How can correlational studies avoid the directionality problem?
Use a longitudinal design
Manipulate the independent variable
Manipulate the dependent variable
Assign participants to different conditions