I think I say you post this chart the other day. It is a damning indictment of generic Congressional polls. Someone else on that same thread posted the following information that may be worth repeating here:
1994 -- DEM +3 Trend Up -- Result R +54 House +9 Senate
1996 -- DEM +6 Trend Down -- Result R -3 House +3 Senate
1998 -- DEM +7 Trend up -- Result R -3 House 0 Senate
2000 -- DEM +5 Trend Up -- Result R -2 House -5 Senate
2002 -- DEM +2 Trend Flat -- Result R +8 House +1 Senate
2004 -- DEM -1 Trend Down -- Result R +3 House +4 Senate
Thanks for putting it in perspective. I posted on the other thread that I had heard pundits say the rule of thumb is the Dims need to be about +7 to gain seats. I sure hope this poll is right.