By Richard S. Dunham

It instantly became part of American political legend: President Bush (news - web sites) received a hefty 44% of the Hispanic vote on Nov. 2, up sharply from the 35% he garnered in 2000. Pundits credited the Latino surge to the GOP -- as measured by two sets of media exit polls -- for Bush's wins in such battleground states as Florida, New Mexico, and Nevada. Sounds impressive. But is it true?
 
Some Democrats and Hispanic activists say it's not. The William C. Velasquez Institute, a nonpartisan think tank, conducted its own exit polls, which found Bush winning 34% of the Latino vote -- no better than four years ago. ``Bad science is being used to misrepresent a community,'' says Robert Aguinaga, research director of the institute. Republicans who derided the media polls on Election Day -- when they showed Democrat John Kerry (news - web sites) running ahead of Bush -- are now embracing them.

In fact, none of the polls appear to get the Latino vote right. BusinessWeek examined real election returns from 62 jurisdictions in 13 states -- mostly places where Hispanics made up 75% to 95% of the population. The bottom line: Bush improved on his 2000 performance in 85% of these heavily Hispanic areas, undercutting Dems' claims that he didn't make inroads. But his gains averaged just three percentage points -- far less than the nine-point gain Republicans have been trumpeting.

Given those relatively small gains, it seems unlikely that Bush could have jumped to 44% of the national Hispanic vote. The President polled just 16% in Philadelphia's overwhelmingly Hispanic 7th Ward, and he took just 12% of the vote in Chicago's 22nd Ward, which is 91% Latino. He did better in suburban areas but usually scored in the low 40s. Bush's greatest appeal was to rural Hispanics: He took 50% or more of the vote in several heavily Latino counties in South Texas and averaged 41% in the counties along the U.S.-Mexico border. And he carried a majority of the Hispanic vote in Florida, although his percentage declined in Cuban-American precincts in Little Havana and Hialeah.

Missing Conservatives

Where did the dueling exit polls go wrong? Velasquez Institute statisticians say the media polls interviewed too many suburbanites. And Democratic pollster Sergio Bendixen posits that the institute focused too much on the barrios. The researchers defend their methodology. ``We stand behind the (44%) number,'' says Warren J. Mitofsky, who performed one of the widely quoted exit polls for the big-media National Elections Pool. Furthermore, Mitofsky suggests that the BusinessWeek study -- which included a cross-section of Hispanic locations, from the inner cities of the North to the rural Southwest, from suburbs to the farm belts of Washington state and Kansas -- missed conservative Hispanics living in predominantly Anglo areas.

Bush advisers say privately they don't believe the President reached 44% of the Latino vote. But that's not stopping them from trying to build on the momentum generated by the post-election buzz. Bush has moved quickly to consolidate any gains among Hispanics.

The message from BusinessWeek's numbers is sobering for Democrats. The party is counting on Hispanic population growth to carry it back to power. But the Dems have lost ground among Latinos in each election since Bill Clinton (news - web sites) won 72% in 1996. To turn things around, Democrats will have to do more than attack exit pollsters.