a1 University of Wisconsin–Madison
a2 University of Virginia
We present an individual-level analysis of the effects of campaign advertising on vote choice in the 1996 Senate elections. Drawing on a unique data set that merges the 1996 National Election Study with detailed satellite tracking data showing when, where, and how many times every Senate campaign ad was aired in each of the nation's top media markets, we develop an alternative measure of media exposure and a different way of measuring the effect of campaign spending. We use this new approach to address long-standing questions about the effect of political advertising and the role of campaign spending in congressional elections. This cleaner measure yields results contrary to some of the conventional wisdom on the effect of campaign spending.
(Online publication January 11 1999)
(Received March 20 2000)