While a GOP wave can truly happen on November 4, it could also very well not happen. The latest Zogby Analytics poll of 898 likely voters shows that 39% of likely voters would choose the Democratic candidate for Congress and 35% would select the Republican candidate, if the election were held today. Another 5% intend to vote for a different candidate and 21% are still undecided. The poll of 898 likely voters nationwide was conducted online October 3-5 and has an overall margin of sampling error of +/-3.3 percentage points.

All of the usual caveats about the "Congressional generic" number prevail. Congress is elected, of course, district by district, and there are dangers in leaning too hard on averages. Nonetheless, generic polls like this have a long history and do reveal some well worn patterns. Normally, Democrats need about a five point lead nationwide to at least break even in an election. Their four point lead is still anemic. Polling in the thirties by both parties with about a month to go is a separate story in itself. But perhaps of greatest significance are the last two numbers - 5% intending to vote for "other" candidates and one in five still undecided. There are so many razor thin races for both the U.S. Senate and the House of Representatives, that a 5% figure can loom very large. Indeed, while there are plenty of Green candidates scattered all over the map, it is the official Libertarian candidates who are polling well enough - at least at this point in time - to hurt GOP candidates in those closely contested races.

But the sheer magnitude of the undecided voters raises plenty of important questions. First among these is that among the 30% who identify themselves as "independents", 14% indicate they will vote for a non-major party candidate and 49% are undecided. Let me repeat: one half of the independents are undecided. Anyone at this point making a "prediction" for how this election cycle will turnout is being very foolhardy. How on earth do we know how many of these self-described "likely voters" will actually turn out to vote?

Just who are these undecided voters? Contrary to the past, they seem to be older. Younger voters lean Democrat (48% Democrat, 24% Republican, 7% other, 21% undecided) among 18-29 year olds; 30-49 year olds (45% Democrat, 32% Republican, 6% other, 17% undecided). Older voters tilt more toward the GOP and are more likely to be undecided): 50-64 year olds (31% Democrat, 42% Republican 3% other, 24% undecided) and over 65 (30% Democrat, 44% Republican, 2% other, and 24% undecided).

Only 8% of Democrats and 11% of Republicans undecided, as are 11% of liberals and 15% of Republicans. Whites lean toward the GOP 43% to 30%, with 22% undecided - not very good news for the Republicans. But perhaps more ominous is the news that Hispanics and African Americans offer to the Democrats - 14% of the former and 18% of the latter are still not sure which party they will support. Among these two groups, undecided voters at this point in time are less likely to vote. If those numbers hold, then this election could be a replica of 2010 - a turnout model that favors a GOP wave.

Thus the conundrum facing Democratic candidates: only two people, Barack and Michelle Obama, have the potential to persuade young Hispanics and African Americans to vote. But these same candidates are trying to stay arms length from POTUS and FLOTUS.