Tuesday, August 5, 2008

Handicapping the Senate election

So, Jim Martin vs. Saxby Chambliss for U.S. Senate. Hopefully it will be more interesting than the primary. I feel sorry for Thomas Wheatley, who had to cover it for Creative Loafing and try his damndest to come up with interesting story lines. That couldn't have been easy.

Chambliss is obviously the favorite. FiveThirtyEight.com uses statistical analysis in addition to polling to project election results, with very accurate results throughout the Democratic primary. The regressions are based on demographic analysis, past voting patterns, etc. Anyway, FiveThirtyEight crams all these polls into a regression analysis and runs 10,000 simulations to predict the possibility of election results. Currently, FiveThirtyEight gives Saxby a 100% chance of winning.

[UPDATE: I believe this regression analysis used Jones' polling numbers, which were significantly worse than Martin's. I expect Martin will have about a 5-10% chance of winning if they update the numbers. Compare Obama's low chance of victory (see below), and the NC Sen regression (a 10 point projected differential) showing a 10% chance of Dem victory. Martin is still a major long shot.]

So, can Jim Martin win? No. I'm not sure that campaigns make a huge difference on elections. Certainly they make a difference, but I think it is usually only during very close elections. I was interviewing for a job on the Mark Taylor for Georgia campaign,* and I asked the campaign consultant, "Can we beat this guy?" His response was, "Well, if the State of Georgia really wants to re-elect Sonny Perdue, there isn't anything we can do about it." Taylor lost by almost 20 points. A better campaign might have made it closer, but Sonny Perdue wasn't going to lose in 2006. I think this election will be similar for Martin. I don't think voters are really that interested in getting rid of Saxby Chambliss.

Can he come close? Maybe. For Martin, the bar is the 42% he got in 2006 vs. Casey Cagle. I think he has a good chance to top it, if only because Obama's presence in the state will help with voter turnout. I think 44%-45% is realistic, and anything over 46% will be incredible. Cleland got 46% of the vote against Chambliss as an incumbent in a very bad year for Dems.

Now Chambliss is the incumbent, and Georgia was one of the few states to buck the national trend favoring Dems in 2006. The rest of the country decided they were done with the GOP, but Georgia doubled down. Sure, Obama may have an outside shot at winning Georgia, but I really doubt it. FWIW, FiveThirtyEight gives Obama a 10% chance of victory.

So what can Jim Martin do make this race more competitive? Five things that could help:

  1. Get a hair cut. This isn't my idea - I've had numerous folks tell me he looks like a used car salesman. Ouch. But still, its true. Appearances matter. Half of Mitt Romney's initial appeal to lots of folks was that he was handsome. And it's not like haircuts haven't been campaign material before.

    Anyway, Martin looks like Gene Talmadge, but without the fiery attitude. Jim's not very exciting, so he just comes off as oily. I think the only way you can pull off that look is by being a firebrand (and I only say that only because Gene Talmadge kept getting re-elected. I always thought he looked oily, too). Which brings me to my next point...

  2. Attack Chambliss. Jim has a reputation as being a nice guy, and it's one reason that this Dem voted for him reluctantly. Voters don't really respond to "nice guy" - especially when the GOP is going to be running on national security. Also, the only way to get close to Chambliss at the polls is the bring him down a notch or two. There is a reason negative campaigning sticks around - it works.

  3. Find a narrative. The Martin campaign never had a coherent story during the primary, other than "electable". During the run-off, it was basically "I never voted for George W. Bush". Martin has to find a 5-second explanation for why he is a better choice than Chambliss. So far all his ads seem like boiler plate Dem territory, and that just isn't enough.

  4. Raise a ton of money. This one is a no-brainer. He should be able to let the Obama campaign take care of voter turnout, which is a rarity for this state, and he should focus his money on TV ads and direct mail.

  5. Don't shoot yourself in the foot. Don't run any more ads talking about how his daughter got kidnapped walking to school when she was 8. It made him look weak, like he couldn't protect his own family.

    Also, he needs to have something ready for when the mailers talking about the deaths at DHR under his watch come out. You know they will, so be ready.
* I ended up working for the Taylor campaign doing research. I spent 4 months on staff and realized I didn't have the stomach for campaigns - literally. The stress sent me to the hospital with a stomach condition, and I quit before the primaries.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.