Time‐lags in Cross‐lag Correlation Studies: A Computer Simulation

ABSTRACT The cross‐lag correlation method has gained increased use in recent years as a technique which can be used to evaluate causality in longitudinal behavioral research. However, the effect of selecting an incorrect time‐lag has remained largely unexplored. This research reports the results of a computer simulation designed to evaluate what happens to the cross‐lag correlation under different time‐lags. The results of the simulation indicate that the cross‐lag method is relatively insensitive to errors in the selection of the time‐lag. Nevertheless, it is better to err on the “early” side than on the “late” side. A research design is suggested for further empirical investigation of the effect of the choice of time‐lags.