What can we learn by examining high profile roll call votes? There are at least three important lessons that can be gleaned from these cases. First, the empirical results contribute the centuries--old philosophical debate about the appropriate relationship between elected representatives and their constituents. We can identify under what conditions citizen monitoring ``turns on,'' and when (and for whom) the switch is off. We can test whether monitoring is continuous or intermittent and predictable or idiosyncratic. Second, respondent knowledge about these votes tell us something about political informedness, across institutions (the Senate and the House) and over information environments (campaign and inter-campaign periods). Finally, these two votes speak to the foreign and domestic policy distinction, and how levels of public awareness might vary across these dimensions.
Why choose high profile votes? The problem with high profile votes is precisely that they are high profile, and thus atypical. However, for a study of citizen knowledge, this is a strength. The typical bill before Congress is subject to little debate, no amendments, and passes into law (or into oblivion) rapidly. Public awareness is certainly low. It is not among these sets of issues that scholars of the Congressional policy--making process (Arnold, 1990) or roll call voting (Kingdon, 1989) focus their lens. Instead, it is the high profile bill which reveals all the twists and turns of the policy making process. For the same reason, we choose high profile votes to gauge constituent knowledge. It is unusual to find circumstances where constituents know the name of their representative, much less the details of a particular vote. Even if infrequent, issues do arise which deeply divide Congress, which are hotly debated in the press, and which intrude upon the public consciousness. It is among this subset of issues that a different gauge of the representative process may be found, where we can determine the boundaries of constituent knowledge of representatives' behavior.
One critical reviewer of our work on the Gulf War vote argues that our
piece should not be the first, but the last paper assessing citizen
knowledge of roll call positions.
The levels of awareness we unearthed were so low,
in this reviewer's opinion, that we mainly demonstrated lack of
knowledge about roll call votes. After all, if a highly salient,
publicly broadcast congressional vote failed to penetrate the
consciousness of 75% of the respondents of a national survey, no
examination of citizen knowledge was relevant outside of the campaign
arena. Democracy can survive even if citizens are only informed during
campaigns and elections.
However, besides needing to convince future reviewers of the merits of our work, we remain unsatisfied with only a single examination of citizen knowledge of roll call votes for other reasons. While some aspects of the Persian Gulf War resolution made it a nice first case --- it was high profile, and members publicly agonized over their positions --- there were other ways in which is was idiosyncratic. First, we would like to vary along the policy dimension, especially looking at issues which are more traditionally seen as part of the Congressional mandate. As extensive research has shown, public opinion about foreign and domestic policy differ in important ways, especially on the perceived legitimacy of congressional involvement. We are particularly interested in the second vote chosen in this study, the Clarence Thomas vote, because it touches on sensitive, and electorally threatening, issues of race and sexual harassment.
Second, we know that information flow is higher during an election year. Some of the low levels of awareness of the Gulf War vote might be attributed to the nature of the times, mid--winter of a year without a federal election (1991). The Thomas vote occurred in late 1991, but the survey assessed recall during the next election year. Conditional on the degree to which Thomas was raised as a campaign issue, public awareness of the vote should be relatively higher.
Third, our previous work unearthed fascinating, though difficult to resolve, differences in constituent knowledge between the House and Senate. In this paper, we control away these differences, and only examine citizen recall of Senators' votes. This should minimize cross--institutional complications, although it does limit to generalizability of our findings. It remains our intent to pursue institutional differences in future work.
To summarize, Alvarez and Gronke (1996) examined the level of constituent knowledge of their representative's votes on a single, albeit highly salient congressional vote, the vote on the resolution to use force in the Persian Gulf. In this piece, we compare citizen knowledge of their Senators' votes on this roll call with another roll call, the vote to confirm Clarence Thomas as a Justice in the Supreme Court. We replicate the Gulf War results from the 1991 survey with 1992 SES data and compare the rates of accuracy with rates for the Thomas vote. In the next section, we describe the two sets of questions we use, and report the basic levels of knowing and guessing for the two votes.