38% D, 28% R, the rest Independents....also stated that they “oversampled” blacks to better represent their percentage in the voting population ( I don’t have a clue as to what this means)....
I assume this means that their sample of adults (or registered voters) included more than 12 percent blacks, which would tend to skew the numbers towards Obama.
This is a bad sample. It isn't representative of the people who will actually vote in this election. In every election since 1988, the percentage of Democratic voters never exceeded the percentage of Republican voters by more than 4 points. In 2004, the Democrats did not have an advantage at all. Here, the spread is a whopping 10 points.
Another problem with this poll is that the likely voter numbers were nearly identical to the registered voter numbers -- yet another indication that the sample was overpopulated with Democrats. According to Gallup, the likely voter pool should be about 4-5 points more Republican than the registered voter pool.