How likely is it that Karl Rove is right, and that Republicans will hold both the House and Senate in the 2006 elections? I don’t pretend to know — and remember, I’m the guy who once again is betting the Philadelphia Eagles will win the Superbowl, but let me offer a few thoughts as to why the GOP might very well have grounds for confidence. Let me also note that I don’t have any particular Rove obsession: He’s human, and he may well be absolutely wrong or saying something he knows is wrong for tactical reasons–declaring “Woe is me” might well be one of those sorts of things that depress turnout. In any case, here goes…

In the 2004 election, the GOP pioneered the use of political datamining and microtargeting–sending individualized messages to voters based on such variables as the magazines they subscribed to or whether or not they ordered college football packages from their cable stations. They identified their likely voters and used messages tailored to the concerns of each voter–family values, low taxes, gun rights, whatever–to get them to the polls. Republicans were able to reach voters directly, intimately, and without using clunky television ads that might energize one constituency while alienating another. A candidate could stress her support for free trade and open borders to Reason subscribers, her pro-life credentials to National Review subscribers, and her support for Israel and democratizing the Middle East to Weekly Standard subscribers, directly, privately, and without the risk of upsetting libertarians with anti-abortion rhetoric or social conservatives with a pro-immigration stance.

The polls look uniformly bad for Republican prospects of holding both Houses of Congress (although at various times the House has looked safer than the Senate, and vice versa), but suppose the GOP had something better than polls: suppose they had a database of people they knew would go to the polls and vote for their candidates? Instead of extrapolating from a weighted sample of 660, a thousand or twelve hundred voters, suppose that, district by district, Republicans have a pretty accurate count of how many votes–total votes–they’re likely to get (and, of course, an estimate of how many they need to win).

They had something like that in 2004. Here’s a bit from a Thomas Edsall and James Grimaldi article that ran in The Washington Post on Dec. 30, 2004:

Republican firms, including TargetPoint Consultants and National Media Inc., delved into commercial databases that pinpointed consumer buying patterns and television-watching habits to unearth such information as Coors beer and bourbon drinkers skewing Republican, brandy and cognac drinkers tilting Democratic; college football TV viewers were more Republican than those who watch professional football; viewers of Fox News were overwhelmingly committed to vote for Bush; homes with telephone caller ID tended to be Republican; people interested in gambling, fashion and theater tended to be Democratic.

Surveys of people on these consumer data lists were then used to determine “anger points” (late-term abortion, trial lawyer fees, estate taxes) that coincided with the Bush agenda for as many as 32 categories of voters, each identifiable by income, magazine subscriptions, favorite television shows and other “flags.” Merging this data, in turn, enabled those running direct mail, precinct walking and phone bank programs to target each voter with a tailored message.

“You used to get a tape-recorded voice of Ronald Reagan telling you how important it was to vote. That was our get-out-the-vote effort,” said Alex Gage, of TargetPoint. Now, he said, calls can be targeted to specific constituencies so that, for example, a “right to life voter” could get a call warning that “if you don’t come out and vote, the number of abortions next year is going to go up. ”

Dowd estimated that, in part through the work of TargetPoint and other research, the Bush campaign and the RNC were able to “quadruple the number” of Republican voters who could be targeted through direct mail, phone banks and knocking on doors.

Democrats had access to similar data files. But the Bush campaign and the RNC were able to make far better use of the data because they had the time and money to conduct repeated field tests in the 2002 and 2003 elections, to finance advanced research on meshing databases with polling information, and to clean up and revise databases that almost invariably contained errors and omissions.

“Very few people understand how much work it takes to get this technology to actually produce political results. We are one election cycle behind them in this area,” said a Democrat who helped coordinate voter contact in the 2004 campaign.

Suppose Replubicans are still a cycle ahead? Predicting elections on the basis of polls might be like assuming that because 400,000 people subscribe to a newspaper, all 400,000 are going to read the lead editorial, rather than skipping to the comics page to read Beetle Bailey, Dilbert or Zippy the Pinhead. Maybe Rove really does have a reason to smile.

Caveat one: this Washington Post story provides some caution in thinking that microtargeting will make that much of a difference in individual races, and suggesting that the polls have some validity. But see also caveat two: this post from Dan Riehl on polls taken two days before the 2002 elections compared with final results in those same races (although this may say more about individual pollsters than it does about the overall reliability of polls).

It’s also worth asking, in an age of increasing media fragmentation, what messages are getting through to individual voters and from what sources–from newspapers, radio, television news, political ads, or from phone calls from phone banks from callers who have some inkling that someone’s top issue is taxes or immigration or the War on Terror or health care.

Bottom line: While nobody ever gets rich betting on the Cubs after the Fourth of July, as the saying goes, and I will never get rich betting on anything–politics being no exception–it would be well worth looking at how much Republicans are spending on TargetPoint Consultants and National Media Inc. in 2006, and how much Democrats are spending on their microtargeting efforts.