I have a monthly data series of commodity spot prices from 1994 to 2010. I have split the data from 1994 to 2002 and 2003 to 2010 due to introduction of some economic process in 2003. now I have taken log of first differece( log(p2/p1)) and found that the two splitted series are stationary. can I do independent t test on these two sample time series, to check if the mean returns have changed after 2003?