WASHINGTON—Karl Rove must read polls through red-colored glasses.
That might explain why two weeks before Election Day he snapped at a reporter who dared suggest that he saw trouble in the polls for the Republicans.
"I'm looking at 68 polls a week," Rove, the president's political strategist, chided a National Public Radio reporter. "You may be looking at four or five public polls a week that talk about attitudes nationally but that do not impact the outcome of individual races.
"I'm looking at all of these . . . and adding them up. I add up to a Republican Senate and Republican House . . .You are entitled to your math and I'm entitled to THE math."
Well, Rove was wrong. His math wasn't THE math. Maybe it was what George W. Bush once called "fuzzy math."
For the rest of us, the numbers had been adding up for more than a year.
Take, for example, McClatchy's reporting as early as May 2005 that the tide had turned against Republicans, that congressional scandals, violence in Iraq and Republican overreaching were creating a toxic political environment like the one that shadowed the Democrats in 1993—the year before they were swept from power.
Or our October 2005 analysis that Republicans feared a wave of anger like the 1994 tsunami that washed the Democrats away. "It's worse today," Republican strategist Frank Luntz said then.
Or reporting from around the country last March that revealed a dispirited Republican base, frustrated and angry over federal spending and illegal immigration.
Then came McClatchy-MSNBC polling of key Senate races this fall, conducted by the respected Mason-Dixon Polling & Research, Inc. While Rove was right that many media polls surveyed the entire country—and missed on-the-ground differences—he seemed to ignore the work by Mason-Dixon, which specializes in state-by-state research.
Our first poll, released Oct. 2, found that Democrats were in position to gain six seats and win control of the Senate. Democrats did not trail in any battleground state; Republicans did not lead in any. Many analysts by then were writing that Democrats might take control of the House of Representatives, but most called the Senate out of reach.
Our second series of polls, released Oct. 24, found Republicans holding narrow leads in Tennessee and Virginia and close in Missouri—suggesting a Southern battleground would decide Senate control.
Our final wave of polls, released Nov. 5, showed a close contest for the Senate, with Republicans gaining in two overlooked states, Montana and Rhode Island, and pulling away to a 12-point lead in Tennessee.
Some readers didn't like that. One compared me to Nazis. Another suggested I should die. None argued with the numbers, just the writer. But the analysis focused on the newest information—growing Republican chances to hold two states widely written off as lost.
Our partner, MSNBC, went farther. Its headline: "Polls: GOP looks likely to keep Senate control."
Our final polls came very close to what happened on Election Day. Eleven of our 12 state-poll results were within the margin of error, which was plus or minus 4 percentage points. That means you can add or subtract the four points for either candidate, for a total possible swing range of eight points. Most of our polls were much more on target than that.
In Arizona, we showed the Republican up by 8 percentage points; he won by 9.
In Maryland, we showed a 3-point lead for the Democrat. He won by 11 points, exactly at the outside edge of our poll's margin of error.
In Missouri, we showed a 1-point lead for the Democrat. She won by 1 point.
In Montana, we showed a tie. The Democrat won by 1 point.
In New Jersey, we showed 7-point lead for the Democrat. He won by 8.
In Ohio, we had the Democrat ahead by 6; he won by 12.
In Pennsylvania, we had the Democrat up by 13; he won by 18.
In Rhode Island, we had the Republican up by 1 point; the Democrat won by 6.
In Tennessee, we had the Republican up by 12; he won by 3—or 1 point outside the margin of error.
In Virginia, we had the Democrat up by 1; he won by less than 1 percentage point.
In Michigan, we had the Democrat with a 16-point lead; she won by 16.
In Washington, we had the Democrat with a 16-point lead; she won by 19.
"We did pretty good," said Mason-Dixon pollster Brad Coker. "The only two states that anybody would call into question would be Rhode Island and Tennessee. That had lot to do with the timing."
He said both polls found a bump in Republican support after Democratic Sen. John Kerry of Massachusetts suggested that U.S. soldiers in Iraq are uneducated. That bump had disappeared by Election Day, and with it the Republicans' last chance to keep the Senate and prove Rove right.