I am running a regression analysis and I´m not sure what data I should use

#1
I wanna do regression of annual Oil prices (Nominal) against annual GDP (REAL) but I´m not sure if i should do regression of absolute values, or percentage growth
 

Miner

TS Contributor
#2
It depends on the question that you are trying to answer. You can do a regression either way (or on standardized values), but the model and the coefficients mean different things. Decide on the question and match the regression to answer that question.
 
#4
It depends on the question that you are trying to answer. You can do a regression either way (or on standardized values), but the model and the coefficients mean different things. Decide on the question and match the regression to answer that question.
Basically i wanna know how much do changes in oil prices affect GDP, and I´m not sure if i should be using red values, or yellow values (picture above), that said yellow values (percentage growth) perform worse (lower correlation overall)
 

Miner

TS Contributor
#5
I would start with scatter plots to determine what relationships might exist, and whether there are any outliers that might influence the correlations for the better or the worse. The plots would also show whether any relationship is linear or not.