How to account for short-term trends in time-series data?

#1
Hi guys,

I would like to find out which of my 10 independent variables (IVs) best accounts for the changes of my dependent variable (DV) at different points of time within the last 50 years.

My problem: In order to explain the long-term trend of my DV over the last 50 years, I used multiple regression. However, I doubt that the most influential IVs (which the regression model suggests) have had equal influence on the DV over the 50 year time period. E.g. IV 1 might have been very influential in the 70s but lost importance in the 80s. Hence, I want to figure out which IV best explains which year or time period. Do you have any ideas (I use SPSS)? My first amateurish thought was to compare the residuals of each IV for each year, but probably I am totally wrong..

Thank you very much in advance for any help!!

Best,


Simon
 
Last edited:
#2
So you want to independent variables (IVs) best accounts for the changes of my dependent variable (DV) at different points of time within the last 50 years.
But you have to specify which are those periods?
Then individually run regression for that period with stepwise regression. Important IV will be selected.