Intention to Vote, Reported Vote, and Validated Vote

Abstract

Pre-election “intention to vote” has come to substitute for “reported vote” or “validated vote” in many studies of turnout. The differences among them have been little studied. In this report, we examine all three jointly in a single U.S. dataset for what we believe is the first time. We find that the same variables tend to be influential for all three measures, but not to the same degree. In particular, intentions are not always translated into actual behavior, so that explanatory coefficients tend to be smaller for validated turnout studies than for vote intention. Reported vote coefficients generally fall between the other two, because some citizens report their actual behavior while others report their noble intentions.