Princeton Election Consortium

A first draft of electoral history. Since 2004

Meta-Margins for control: House D+1.0% Senate R+4.2% Find key elections near you!

Race and voting

November 25th, 2008, 8:23am by Sam Wang

Courtesy of reader FZ, an excellent map at the NYT. By hitting “voting shifts” you get a mapped county version of shifts not only comapred with 2004, but also earlier years. He also notes analysis by Chicago (formerly Princeton) political scientist Eric Oliver, who suggests that the racial element occurs in islands of white communities in black-dominated surroundings. He suggests racial competition rather than racism per se. Worth a read.

Tags: 2008 Election

One Comment so far ↓

  • Richard

    The Good News:

    The Final Election Model projected an expected 365.3 EV for Obama.

    2008 Election Model

    A Monte Carlo Electoral Vote Simulation

    Updated: Nov. 3, 2008



    Obama wins by 76-64m votes; 367-171 EV (median); 365.3 expected; 53-45% vote share

    The Election Model (EM) assumes as a base case that a fraud-free election is held today – and that current polls reflect the true vote. The model projects that Obama will win the Electoral vote by 367-171 and the True Vote by 76-64m. The final projected vote share is Obama 53.1- McCain 44.9%- Other 2.0%. The state poll aggregate vote share matched the national average tracking poll to within 0.2%.

    In May, the 2008 Election Calculator (EC) projected that Obama would win the True Vote by 71-59m (54.1-44.7%).

    For the 2008 EC to match the EM, its estimate of returning 60.5m returning Kerry and 51.6m Bush voters had to be accurate.

    The EC used 12:22am 2004 NEP vote shares to calculate the projections.

    In other words, the 2008 EC and EM confirmed that Kerry won a landslide (see below).

    The bad news:
    We are both wrong. We had the Recorded Vote right but not the True Vote

    Obama had to have done better than the 365EV and 53%.

    Here’s why:

Leave a Comment