The stochastic model for the evolution of preferences proposed and tested by Falmagne, Regenwetter and Grofman (1997) and Regenwetter, Falmagne and Grofman (1999), as well as an alternative model proposed by Bockenholt (2002), give a good statistical account of attitudinal panel data from the 1992 U.S. presidential election. We show, however, that both models have the defect of underestimating the consistency of some respondents across different polls.
We present a generalization of Falmagne, Regenwetter and Grofman's model based on the idea that some individuals may become momentarily impervious to all matters related to the campaign and 'tune out.' This could be due to some personal reason or to some external event related to the campaign. Like the original model, the resulting model is a random walk, but on an augmented set of states. A respondent in an 'active' state behaves as in the previous model, except if receiving a 'tune-out' token, which effectively freezes the respondent's preference state until it is reversed by a 'tune-in' token. We describe and successfully test the new model on the same 1992 National Election Study panel data as did Bockenholt (2002), Regenwetter et al. (1999).