PPT Slide
Two methods for determining whether a correlation
between X and Y found in observational research
reflects a “forward” cause-effect pattern between X and Y:
If X is experimentally manipulated before Y is measured
(thus eliminating backward causation) and all
other variables are controlled (thus eliminating third
variable causation), then any significant X-Y covariation
must be due to forward X-Y causation.
2. Use advanced correlational procedures.
Partial correlation and cross-lagged panel correlation
measure the correlation between X and Y with third
variables statistically controlled or with X and Y arranged
in either “forward” or “backward” order.