Let's assume that the early votes make up half of the total runoff turnout. If that's the case, then turnout will have dropped by almost 75% from the November 4 total. My guess is that that is a bit too steep a drop off. But I don't think that something between that mark and the 45% figure from 16 years ago is a stretch. Much depends on the each of the candidates' ground games. My observation has been that Chambliss has been on the air more (at least in Clarke County via the Atlanta TV market). This was confirmed this morning by my less than representative American government class, which while not representative from a sampling standpoint was at least dispersed throughout Georgia during UGA's weeklong Thanksgiving break.
That, of course, is only part of the equation though. The GOTV efforts on the ground in the Peach state are where the real difference is going to be felt. There is at least some anecdotal evidence that Martin has marshaled a superior ground game to Chambliss based on the Obama infrastructure and an influx of Obama volunteers to the state. But as you can see below, those efforts have not made a dent in one of Martin's vital voting blocs (...at least not in early voting).
The proportion of African Americans coming out to vote early barely varied throughout the early voting period. From November 14 - 26, that proportion only broke 23% twice; the day that early voting began in Fulton County (Atlanta) on the 18th and on Friday the 21st. And even then, that day's total of African Americans only amounted to 24% of the day's early voters. Even that is far below the over 34% of the early voters African Americans comprised ahead of the November 4 general election. The extent to which Martin's (Obama's) supporters can reach that segment of the electorate prior to 7pm Tuesday night will to a large degree determine whether the former state rep can overcome the three point deficit from November 4 (...not to mention the 3-6 point lead the handful of polls conducted have shown Chambliss enjoying).
One thing Martin may be able to lean on is the steady rise of women early voters across the runoff's early voting period. Though that demographic, too, fell below the mark it reached in the general election early voting, it was, at the close of early voting, only two points below the general election level (54% for the runoff to 56% in the general). That may bolster Martin's numbers, but it could also be that those female voters are coming from Chambliss areas. It may not be a coincidence that the rise in women voters came at the same time as the Cobb County (a Republican stronghold) rose to the top of the pack in terms of raw numbers of early voters county-by-county.
If we break women early voters down by racial groups, we don't gain too terribly much more information. Both demographics increased subtly over the three days in which early voting was conducted prior to Thanksgiving last week. Though white women made up anywhere from 34% of daily early voters to 39%. Black women, on the other hand, varied from just shy of 12% to just north of 14% on any of the ten days early voting was held.
Much of that speaks volumes of the current state of the race. Chambliss looks to be in good position heading into Tuesday's vote, but whether that is a comfortable margin come Tuesday night will depend on the GOTV efforts being made throughout the state today and tomorrow.
Georgia Senate Runoff: Early Voting (Day 7)
Georgia Senate Runoff: Early Voting (Day 6)
Georgia Senate Runoff: The Polls