It's difficult when deciphering polls to ward off wishful thinking. Are young voters without land lines really a stealth bloc that's going to tip the race to Kerry? Are the likely voter numbers favoring Bush really a result of obsolete formulas for gauging commitment? Can Democrats really count on previously apathetic nonvoters? After all, as Democratic strategist James Carville famously said, "You know what they call a candidate who's counting on a lot of new voters? A loser." But Tony Fabrizio, a Republican who served as chief pollster for Bob Dole's '96 presidential campaign, doesn't have much incentive to game the numbers to hearten anxious Democrats. And if his latest analysis is correct, the only way Bush can win will be if fewer minorities turn out this year than they did in 2000, when the stakes were far lower. Fabrizio's latest poll of 12 battleground states shows the race dead even, with Bush getting 47.3 percent and Kerry getting 47.1. But a press release for the survey says: " W hen the data is weighted to reflect minority turnout based on the 2000 exit polls, Sen. Kerry leads by 3.5 percent and if minority turnout is weighted to census levels Sen. Kerry's lead expands to 5.2 percent." "It is clear that minority turnout is a wildcard in this race and represents a huge upside for Sen. Kerry and a considerable challenge for the President's campaign," Fabrizio is quoted saying. "If one assumes minority turnout exceeds their 2000 election levels, then it appears a number of these states would tip to Sen. Kerry." Given the enduring anger of widespread minority disenfranchisement four years ago and the massive get-out-the-vote machines being deployed in swing states, it's hard to imagine that minority turnout will be lower than it was in 2000.