I am testing the lead-lag relationship among two variables with cross-correlation function (CCF) by SPSS. however, SPSS computes CCF coeficient for all defined lags, but it does not determine the p-value for them, how I can obtain the p-value of various lags?

For some variables that are not normally distributed, I used their ranked form in CCF, to simulate a kind of CCF with Spearman's method, in this case, how would it be possible to obtain p-value for each lag?

  • 1
    $\begingroup$ SPSS still computes standard errors (based on the assumption that the series are not cross correlated and one of the series is white noise) and shows corresponding confidence intervals. $\endgroup$ – ttnphns Mar 1 '13 at 5:12
  • $\begingroup$ Yes, but in the time of reporting the result of CCF in the text of manuscript, p-value of CCF is needed. Am I wrong about the way that a CCF should be reported? $\endgroup$ – Idin Mar 1 '13 at 7:21