As discussed earlier, the low-population Red states in the Western/Mountain areas --
while they send a fair number of Democrats to the Senate these days -- vote Republican
in presidential elections. That's important, because electoral votes are allocated by giving
states one vote for each member of the House they send (which correlates roughly to a
state's population size) plus one for each senator. So a state like South Dakota, which
as only one representative but two senators, gets three electoral votes. That gives states
with small populations an edge, and most of the states with small populations go to the GOP.
If you strip each state of its two Senate electoral votes, and add up the 30 states George W.
Bush and the 21 (including D.C.) Al Gore won in 2000 (awarding Florida to Bush), guess who
wins? Gore, with 224 electoral votes to Bush's 211.
The question of how the Electoral College biases outcomes is old hat among political scientists-- and it has two components that can't be taken in isolation from each other. One is the one Tapped notes-- the massive increase in the relative voting power of small states because every state (large or small) gets two more electoral votes than its number of Representatives. The other is the fact (not dictated by the Constitution but by the state law of 48 states) that electoral votes are awarded on a winner-take-all basis by state. Notice that 224-211 is a much bigger margin than Gore's popular-vote plurality (c. 6% vs. <1%). That's driven by awarding all of the electoral votes from California, New York, New Jersey, Pennsylvania, Michigan, and Illinois to Gore. (Bush won fewer big or very big states-- Florida, Texas, and Ohio. IIRC.) Given that, currently, cities vote Democratic and rural counties tend to vote Republican, the electoral college over-rewards Democrats for their urban majorities in states such as New York, and denies Republicans any benefit from their large rural votes in such places.
The Electoral College increases the proportionate weight of very small states; but also creates massive returns to even tiny pluralities in very big states. One overweights for rural constituencies, one for urban constituencies. They don't perfectly cancel out; but they come closer to cancelling out than Tapped's math suggests.