Data Availability Determines Whether Campaigns Focus On The Middle Or The Base

Authored By 
Eitan Hersh
Blog contributor 
Resident Faculty Fellow
August 11, 2015

This piece first appeared on Five Thirty Eight on August 10, 2015. It is published here in its entirety with the author's permission.

The primary season is here, the first Republican debates are in the books, and the presidential candidates are trying to persuade voters to support them. Persuasion happens a lot in primaries, but enjoy it while it lasts: Once campaigning for the general election begins, persuasion will mostly disappear, and mobilization — getting people who already support you to vote — will become the main strategic imperative.


Follow the data.

In elections between Republicans and Democrats, campaigns typically have access to data that makes mobilizing the base easy. In most of the country, the government simply tells political campaigns who their supporters are. The state or local election authority provides a list to campaigns with each voter’s political affiliation.

Since about 90 percent of registered Democrats and Republicans vote for their party’s candidates across offices, the strategy is to identify these partisans and make sure they vote.

But in primary elections, campaigns do not have data to easily identify their supporters. Sometimes, a candidate is favored among an age group or gender within a party, which allows for a mobilization strategy. But mostly, candidates are fighting over primary voters who are difficult to distinguish from one another. Campaign records do not identify voters as devout Christians or secular, sympathetic to libertarianism or socialism, or backers of specific issue positions.

Based on the electorate’s prior behavior, campaigns can anticipate who will vote, but they have poor signals about where voters stand.

In some places, even general election campaigns operate in more of a persuasion world. For example, in Ohio, Virginia and Wisconsin, the state governments do not collect party affiliation data from voters. And campaigning differs in these states. In 2012, Harvard’s Ryan Enos and I partnered with the Obama campaign to study its staff and volunteers. (More details about our study are available here, here and here.) We found that compared with workers in the other swing states, Obama workers in states without party data focused less on mobilization and more on persuasion. Canvassers and phone callers had lots of conversations with voters who were undecided or inclined against their candidate. These persuasion-oriented conversations were not usually intentional. Lacking data that outs voters as Democrats and Republicans, the campaign strategy was simply less efficient than it was in other places.

There is a broader lesson here: Campaign strategies are dependent on data, and data is, in part, dependent on government policy. Small changes in data laws can have a substantial effect on campaigns. If you think our democracy would be better served by general election campaigns focused on persuading voters in the middle rather than mobilizing hard-core partisans, consider this: If state governments stopped sharing so much data about voters’ partisanship with campaigns, campaigns would focus less on drumming up their base and focus more on persuading voters on the issues, and they would be forced to have more frequent contact with voters who generally disagree with them. In short, the general election strategy would look more like the primary strategy.