I use variation across markets in the timing of television's introduction to identify its impact on voter turnout. The estimated effect is significantly negative, accounting for between a quarter and a half of the total decline in turnout since the 1950s. I argue that substitution away from...