Nbadan
11-01-2008, 06:55 PM
The so-called Bradley effect is being invoked as a reason to distrust the polls that show Obama is clearly ahead. The New Hampshire primary is constantly mentioned by the pundits as proof that it is real. Is that true or was it just bad polling? If there really is this Bradley effect it should have shown across the board in the Democratic primary races as a negative for Obama. So did it? Look at data for six of the most competitive Democratic primaries to see if Obama consistently under performed his final polling numbers. Not only did he not under perform he actually slightly over performed....
Data gathered from RCP for the primaries in NH, VA, OH, PA, NC and IN and looked at the final poll versus the actual results.
State Poll Actual Delta...
NH O+8.3 C+2.6 C+10.9
IN C+5.0 C+1.4 O+3.6
VA O+17.7 O+28.2 O+10.5
PA C+6.1 C+9.2 C+3.1
OH C+7.1 C+10.1 C+3.0
NC O+8.0 O+14.7 O+6.7
So Clinton beat the polls in 3 states by an average of 5.7 points.
Obama beat the polls in 3 states by an average of of 6.9 points.
In other words, in 6 key states Obama out performed the poll numbers more than did Hillary Clinton. Where was the "Bradley effect"? It didn't show up because it is a complete myth. If you look at New Hampshire closely you will see that Clinton led consistently in the weeks before the election until a group of polls at the very end showed an Obama surge. This turned out to be incorrect for various reasons. The primary in New Hampshire was more complicated to poll because independents could vote in either the Democratic or Republican primary. One theory is that that many independents switch and voted for McCain when they thought Obama was going to win anyway. In 4 of the remaining 5 states the poll leader won and by more than the final margin in the polls.
Data gathered from RCP for the primaries in NH, VA, OH, PA, NC and IN and looked at the final poll versus the actual results.
State Poll Actual Delta...
NH O+8.3 C+2.6 C+10.9
IN C+5.0 C+1.4 O+3.6
VA O+17.7 O+28.2 O+10.5
PA C+6.1 C+9.2 C+3.1
OH C+7.1 C+10.1 C+3.0
NC O+8.0 O+14.7 O+6.7
So Clinton beat the polls in 3 states by an average of 5.7 points.
Obama beat the polls in 3 states by an average of of 6.9 points.
In other words, in 6 key states Obama out performed the poll numbers more than did Hillary Clinton. Where was the "Bradley effect"? It didn't show up because it is a complete myth. If you look at New Hampshire closely you will see that Clinton led consistently in the weeks before the election until a group of polls at the very end showed an Obama surge. This turned out to be incorrect for various reasons. The primary in New Hampshire was more complicated to poll because independents could vote in either the Democratic or Republican primary. One theory is that that many independents switch and voted for McCain when they thought Obama was going to win anyway. In 4 of the remaining 5 states the poll leader won and by more than the final margin in the polls.