Oct 28, 2010 0
As Election Day draws near and polls put Martin O’Malley’s lead over Robert L. Ehrlich Jr. at five, eight, 10, 11 and 14 points, count on the New York Times for a little clarity.
The statistical gurus who run www.fivethirtyeight.com, compiling mountains of polling data in months leading up to the 2008 presidential election, have migrated their work to the Times and are keeping an eye on hundreds of midterm races. Among those, lucky for us, is the Maryland governor’s race.
They input all the publicly available data into their statistical model and spit out projected election results (O’Malley 53.7 percent, Ehrlich 43.7 percent on Oct. 26) and the chances that O’Malley hangs on to his seat (96.3 percent). A set of forecasts is listed at the bottom of the page.
Look at the data below the charts for the polls they use, and the weight assigned to each in the model. The Oct. 26 model likes the ABC/Post poll that showed O’Malley with a 54-40 lead the best.
Ehrlich has dismissed the recent poll results, saying that an internal poll conducted by his campaign shows O’Malley only up 3 points.
The trend in the race the statistical model reveals is interesting. The race was closest in the middle of July, with just a few fractions of a percentage point separating the candidates. But after that, O’Malley developed a small lead through the rest of the summer. He was the only candidate advertising at that point, using his sizable money advantage to buy up TV and radio spots.
O’Malley’s lead widened in September, when his attack ads began to hit. And in October, O’Malley’s lead really opened up, as individual polls have shown.