>I known it, but when you group it on year basis, the statistical difference is zero.

It depends on how big your set is. You can not say this for a small set of data, so it's better to be accurate <g>