According to a revealing poll from Quinnipiac University (covering 2,181 registered voters in late July), only 36% of American voters would support Mr. Obama against an unnamed Republican candidate "if the 2012 election were held today." The main reason for the president's performance in this survey is his pathetic standing among self-identified white voters: Only 28% of the nation's demographically dominant racial group plans to back him for a second term.

Republicans look at those numbers and say there is no way that Mr. Obama can recover without bringing about a major turnaround with the white majority. Yet Democrats point to the figures and argue that the president will safely win a second term even with this dismal performance in the white community—as long as he replicates his 2008 popularity among African-Americans, Latinos and Asians. They also believe he may do even better among Latinos and Asians when he runs in 2012.

Is this reasoning realistic or ridiculous?

The truth is that Mr. Obama's low standing among white voters is nothing new. He lost that group to John McCain in 2008, winning only 43%. If he fails to improve his terrible standing in the current Quinnipiac poll, and if all currently undecided white voters (25%) break down in the same way as those who have already made up their minds, he'd end up with 38% of white votes.

That's obviously a worse performance than four years ago, but it would yield approximately the same percentage of the overall electorate. Why? Because all observers agree that white voters will comprise a smaller piece of the total voting population than the 74% they represented two years ago. With strong increases in the Latino and Asian voting blocs—due to general population growth and sharply increasing rates of citizenship through naturalization—the "non-Hispanic white" electorate will likely slip to 70%, or perhaps slightly lower.

If the president performs as poorly in the white community as current polls indicate, he will still win an electoral majority as long as he commands the same percentage of nonwhite voters (83%) that he won in 2008. This seems entirely possible, and based on current polls, it looks likely.

The Quinnipiac survey indicates that Mr. Obama still enjoys huge popularity among people of color, winning his trial heat against an unspecified Republican 44 to 1 among blacks (87% to 2%) and nearly 2 to 1 among Latinos (49% to 26%). In other words, the president maintains his near unanimous support in the black community and has dipped only slightly among Hispanics, where he drew a commanding 67% of the vote in 2008.

Only 65% of Latino voters expressed a candidate preference in the survey's trial heat. That means if Mr. Obama can sway the bulk of the 35% of Latinos who say they "don't know" or are currently uncommitted, the president will replicate his victory formula from 2008. Undecided Hispanic citizens, representing as many as three million votes in the next election, may hold the balance of power in a competitive race.

These numbers help to explain the president's current position on immigration reform and his efforts to block Arizona's tough new immigration law. That legislation is overwhelmingly resented among Latino voters: 66% of Hispanics say they disapprove of it, and 71% say they don't want a similar law in their own states. By nearly 2 to 1 (59% to 32%), these Latino voters want immigration reform to emphasize "integrating illegal immigrants into American society" over "stricter enforcement." This is in stark contrast to both white voters and black voters, who strongly prefer "stricter enforcement."

The administration and its strategists reason that nothing they do on illegal immigration will undermine the enthusiastic support for the president in the black community, or drive his popularity lower among whites. With only 28% of white voters currently committed to backing Mr. Obama for re-election, his standing is already near rock bottom.

But if the Democrats can use the immigration debate to drive the president's numbers even higher among Latinos than in 2008, they can't lose. Viewed another way, if Republicans continue to conduct the immigration debate in a way that drives their numbers even lower among Latinos than in 2008, they can't win. Talking about changing the Constitution to eliminate birthright citizenship, for instance, may bring short-term gains, but it will produce disastrous long-term results in the key voting bloc that is likely to decide the next presidential race.

For Democrats, this analysis offers reason for hope in a dark, dysfunctional season. For Republicans, the numbers suggest a strong basis for recalibration as they look ahead to 2012.

**Mr. Medved is the host of a daily, nationally syndicated talk radio show and author of "The 5 Big Lies About American Business" (Crown Forum, 2009).