Now that we’re in the middle of March Madness and nearing the opening of the 2015 Major League baseball season, we see the sports data geeks take center stage. Ever since the book, and subsequent movie, “Moneyball,” fans have been intrigued by the data that appears to be a major driver in sporting decisions, from the players chosen for a team, to the offensive and defensive formations, where and when the percentages suggest shooting or passing, and so on.
Fantasy Football, and other online games where player stats are used to allow friends and co-workers to make custom teams and compete against each other, are another example of how computers and data have transformed the sporting experience.
Most campaign data, like who a candidate mailed to, who received phone calls or door knocks, who was identified as a supporter… those are all factors that are not made public and cannot be analyzed.
But, for those in politics, it begs the question: When will campaigns have their “Moneyball” moment?
Sports provides statisticians with an amazing bounty of perfectly preserved data, and millions of datapoints that can be used and manipulated to identify key factors that are not visible to the more casual sports observer. Each game is like a single laboratory experiment that is fully public with the ability to draw down data on who played, the location, elevation, and particulars of every field and playing surface, the length of every pass thrown or speed of every strikeout pitch.
Campaigns, in contrast are data-poor. There are some variables that can be used for analysis, like campaign spending, basic ethnic and gender information on the candidates, the partisanship and demographics of the electorate, and who voted on Election Day. But most of the campaign data, like who a candidate mailed to, who received phone calls or door knocks, who was identified as a supporter… those are all factors that are not made public and cannot be analyzed. Same can be said for campaign messaging. Data on campaign themes or content of speeches is available for a Presidential campaign, but not for your local legislative, city council or school board race.
To use the comparison to sports, the data available for most political analysis would be the equivalent of having the demographics of the teams, players, the attendance, the direction the wind is blowing, and the final score, but none of the specifics about how the game was won.
Does this mean that data doesn’t matter? Far from it. Experts in the field may have a fraction of the overall data, but they are working at a faster pace than ever to make it work for candidates and consultants.
There is an explosion in the use of data modeling to take polling results and apply scores to every voter in the state. Campaigns big and small are developing methods to better track the impacts of communication with voters, whether it is traditional mail, cable television buys, or online advertising. Voters are getting cookie matched and micro-targeted to the gills.
Some saavy associations and PACs are building data on voters that can be carried forward from campaign to campaign. Instead of disposing of the mountain of data built up for a campaign one year, then starting from scratch for the next, they are making use of the lessons learned about particular voter behavior and attitudes, and using it for multiple campaigns. And even local candidates and campaigns are learning the benefits of a long-term data collection and use of political databases to track and communicate with voters in the intervening years between elections.
But will there be an algorithm to help you with your 2016 Primary Election Pool? Don’t hold your breath.
Ed’s Note: Paul Mitchell, a regular contributor to Capitol Weekly, is a political strategist, analyst of campaign data and the owner of Redistricting Partners.