Proposal for the 2016 Time Series: Propensity-To-Vote (PTV) items

We propose the inclusion into the ANES of a small battery of PTV (propensity-to-vote) items. First introduced in the Netherlands in 1982, such items have since then become routinely employed in a growing number of countries (including, on several occasions, the U.S.). They are aimed at tapping into voter preferences and utilities associated by the respondent to each of the available political parties; as such, while not particularly expensive in terms of questionnaire space, they have proved extremely productive – in a number of countries through time – for studying party preferences, especially in terms of how such preferences overlap across different parties; they also yield a specific additional potential for studying partisanship in the U.S.

Read the full proposal

One thought on “
Proposal for the 2016 Time Series: Propensity-To-Vote (PTV) items

  1. Mark Franklin

    I have used PTV variables extensively in cross-national voting research. In addition to the advantages and opportunities mentioned in the proposal, let me just mention two more topics for which PTV questions have been found useful outside the US and might be found equally useful by US scholars: negative partisanship (since low scores on PTV variables amount to declarations of negative partisanship regarding the parties concerned) and the impact of party preferences on turnout (someone for whom no party is attractive is much less likely to vote, providing a much needed link between party preferences and turnout). Indeed there are a great many topics where PTVs can provide insights and controls, just as in these cases, reducing the need for topic-specific questions.
    One thing cannot be stressed too much. These questions get their power from the fact that they are asked about each party, and respondents’ answers regarding one party in no way constrain their answers regarding another party. US scholars often find it hard to grasp how such questions make sense, so used are they to questions that divide respondents between Dems and Reps such that the answer given in regard to one party constrains the answer that can be given regarding the other. The PTV battery escapes this constraint by focusing respondents’ attention on the longer term, over which it is logically possible (and happens extensively) that a citizen can vote for different parties at different elections. It is not at all unreasonable in the context suggested by these questions for a respondent to tell us that they are very likely to “ever vote” for both a Republican and a Democrat (or never to vote for either).

Comments are closed.