Too funny. The Democrats helped defeat Eric Cantor not so much by voting against him in the Republican primary as by handing the Flat Earth Dave Brat-supporting insurgents some science.

In primary campaigns, the normal procedure is to ignore voters who don't have a history of voting in primaries and focus first on those who do. This is the opposite of the strategy for general elections, where those who vote in primaries are ignored because they will presumably get themselves to the polls. To beat someone like Eric Cantor in a primary, however, it would be necessary to mobilize unlikely voters. The Tea Partiers didn't know how to do that, but the campaign manager and political director of Eric Holder's 2010 opponent did.

Here's the scoop from the campaign manager, Brian Umana:

The tea partiers already knew how to mobilize the folks who showed up at tea party meetings: what they needed was a way to find supporters or potential supporters who were unlikely to bother with regular meetings. [Political Director Jonathan] Stevens and I thought that a more organized attack from the right could help Democrats, too—either by prompting a future three-candidate race (which might give the Democrat a fighting chance) or by inducing a competitive Republican primary challenge that would force Cantor to burn cash protecting up his flank that might otherwise be spent on competitive races elsewhere. (A primary campaign resulting in Cantor’s defeat, of course, hardly crossed our minds. When [future Brat consultant Tammy] Parada mentioned it, I recall calling the possibility “fanciful.”) Stevens and I saw no harm in mentioning strategies that tea partiers might use to reach sporadic Republicans or far-right “independents” who were less likely to support Cantor than other Republicans. We shared data-science techniques for voter targeting and for evaluating the relative cost of earning the votes of different types of voters.

There was a problem: the-easiest-to-use political data is owned by the two major political parties. The Democratic campaign was over, so how could we ethically share information that we thought would serve the greater good? Stevens used his statistical knowledge and near-photographic memory to work from crude, publicly available State Board of Elections data, then manipulate those data into targeted sets of voters more like those that would be available to a large campaign from one of the two parties. He created tidy data sets of voter information and preferences of a sort typically unavailable to independent or insurgent campaigns opposed by a party establishment (like Mr. Brat’s this year). Some techniques like Stevens’s had been used by Obama’s presidential campaign—which Stevens worked on in 2008—but they had not been widely adopted by Republicans, let alone tea partiers without access to the big party databases. Now Parada, who was at our post-election meetings in 2010, knew how to use them.

...

As Nate Cohn proved fairly decisively by looking at precinct-level data, the Democrats that voted for Brat were statistically significant but very unlikely to have provided the margin of victory. Cantor got waxed because the Tea Party had science.

http://www.boomantribune.com/story/2014/6/14/9552/40093