Eitan Hersh’s book proceeds from a simple but profound insight: electoral campaigns go to war with the data they have, not the data they might want or the data we might imagine. Whereas many other studies of electoral behavior rely on secondary indicators to approximate what campaigns do, Hersh delves into the actual databases that (Democratic) campaigns rely on in their work. The result is a rich, detailed, and insightful account that challenges much of what we previously believed about how campaigns make use of voter data in elections. This is a benchmark text, one that I expect will have a long-standing impact on the direction of the field.

The core of Hersh’s work is what he calls the Perceived Voter Model. “Perceived voters compose the electorate from the campaign’s-eye-view. They are not people; they are avatars generated from whatever data a political campaign, candidate, or party can surmise . . . Campaigns do not perceive voters as voters perceive themselves” (pp. 7–8). The research literature is filled with studies that either assume campaigns to be rational actors with near-complete information or seek to understand campaigns through post hoc analysis of what voters recalled. Campaigns are not omniscient, though, and the promise of “big data” proves to be something of a mirage in pract