There's a lot of uncertainty surrounding the state of the election.  Nearly every analyst concedes that the Republicans will gain seats, but the debate surrounds whether it will be enough for them to gain control of the House and Senate.  Many are admitting (Bob Shrum excluded) that the House is a very likely loss for the Democrats, but most picture a scenario of a 40-45 seat turnover with a ceiling of the mid 50's.  Even fewer believe the Republicans are likely to gain the ten seats needed to regain control of the Senate, holding that a 6 to 8 seat gain is more likely. 

Guiding a lot of these projections are traditional understandings of the demography of the nation and the polling data.  Recent polls have sent alarm bells among Democrat camps as generic ballots have turned Republican.  The "enthusiasm" gap has also been disturbing for many of the Democrat pundits.  Despite this, there has been a general lessening of anxiety as several of the polls have fluctuated a bit in their favor. 

The general argument goes that Republicans are destined to pick up some seats as the Democrats have (A) overextended their hold into deeply Republican districts in recent elections, (B) it's a mid-term election which often hurts the party in power anyway, but that (C) gains will be limited to cherry-picked districts. 

A recent poll from an obscure district in Michigan provides interesting insight into why most of the experts are wrong to relax. 

The Michigan 15th district is a seat held by Rep. John Dingell.  He's currently the longest serving member of the House and prior to him his father held the seat.  A Dingell has been in Congress since 1932.  Dingell represents a district that is solidly Democrat.  According to the Cook Partisan index, this district is a D+13, which means that a Democrat will average 13 points above the national average (which is an additional two points in the Democrat favor).  So just to break even, a Republican challenger needs to gain fifteen points on his opponent. 

So what does this recent poll show?  Political newcomer Rob Steele is leading John Dingell by 4 points.  If this holds true, this represents nearly a 20 point Republican swing!  For perspective, consider that there are 68 Democrats in seats with a Republican favored Cook Partisan rating.  If you add all the D+13 and under seats there's an additional 110 seats in play.  If Dingell is an entrenched Democrat in a solidly Democrat leaning district and he's polling under his challenger, how many other races are in the same boat?

So what's going on here?  In early polling, people are more likely to remain undecided unless they know the candidate (which greatly helps the incumbent).  As a general rule, the primary indicator in early polling is not how close the race is polling, but rather if the incumbent is polling above 50%.  Below the 50% mark indicates trouble for the incumbent.  As the election gets closer, the greater indicator is how many undecided voters remain undecided.  If the incumbent is still leading by a fair margin and there are very few undecided voters, then this is also good news for the incumbent. 

The best indicator down the stretch is to see how the independents are voting.  If it's as bad as the polls are showing, as Jay Cost of the Weekly Standard highlights, it's extremely bad news for the Democrats in November.  Just how bad?  Let's just say that I'm sticking to my prediction of a 79 seat turnover in the House and 10 seats in the Senate, but if the Michigan poll is an indication of things to come, it could be well over 100 and 12.