ETM - lagged linear correlation



I am writing because some confusing has arrised in my classes when using the lagged regression in the linear modelling panel. 

lets say that I have an ENSO index file as the independant variable and a time series file of NDVI (monthly mean) as a dependant variable. 

My goal is to find out how long after an ENSO event the highest correlation is found. in total i create 12 different files each containing the R value for each pixel at a lag from 1-12. After that i determine at which lag the R value for each cell was highest. 
The confusion arrise when entering the lag in the panel. 

According to the help menu the lag should be entered as negative number, e.g.  "-3", however in your tutorial video it is said that the lag should be entered as a positive number, e.g.  "3".

So I was wondering if anyone could help clarify this problem? 
From my data (which is not actually NDVI) I cannot figure out which one is correct by solely looking at it. 

Cheers, Beate


Please sign in to leave a comment.