Measuring Voters' Knowledge of Political News

We propose a methodology to measure knowledge of news about recent political events that combines a protocol for identifying stories, a quiz to elicit knowledge, and the estimation of a model of individual knowledge that includes difficulty, partisanship, and memory decay. We focus on news about the Federal Government in a monthly sample of 1,000 US voters repeated 11 times. People in the most informed tercile are 97% more likely than people in the bottom tercile to know the main story of the month. We document large inequalities across socioeconomic groups, with the best-informed group over 14 percentage points more likely to know the typical story compared to the least-informed group. Voters are 10-30% less likely to know stories unfavorable to their political party. Also, each month passing lowers the probability of knowing a story by 3-4 percentage points. We repeat our study on news about the Democratic Party primaries