Academic journal article Presidential Studies Quarterly

Measuring Congressional Support for the President: Evaluating NOMINATE Scores

Academic journal article Presidential Studies Quarterly

Measuring Congressional Support for the President: Evaluating NOMINATE Scores

Article excerpt

Scholars of both the presidency and the Congress require an accurate measure of legislators' underlying preferences. Whether one is interested in explaining policy outcomes, the politics of congressional committees or parties, or bargaining between the White House and Congress, one must know the preferences of the relevant actors. We cannot say anything about presidential influence, then, without controlling for the other determinants of legislators' actions, such as personal preferences and ideology or concerns induced on them by constituents and interest groups.

Some research on legislative-executive relations has focused on improving measures of the extent to which Congress supports bills favored by the president (e.g., Edwards 1985; Fleisher and Bond 1983). These studies use data typically derived from congressional voting, preferring to filter (or censor) the information they obtain by selecting the "most appropriate" type of votes for their analysis. As an alternative approach, Poole and Rosenthal (1985, 1987, 1997) have developed a new approach to recovering the positions of legislators (ideal points) and of proposals in a policy space.(1) This article evaluates Poole and Rosenthal's approach, called NOMINATE, by assessing the degree to which it makes accurate predictions about the positions of members. It concludes that despite important advantages, the NOMINATE techniques seem to generate significant errors in identifying the positions of members. Section two of this article describes the NOMINATE approach emphasizing its advantages and assumptions. In section three, we describe our testing procedures. Section four summarizes our results suggesting that NOMINATE makes some serious errors. The final section suggests that NOMINATE's problems are generic to a class of approaches, including the use of Americans for Democratic Action (ADA) scores and most other voting scores. It also suggests two new approaches to analyzing member behavior to establish the baseline behavior necessary for judging presidential influence.

NOMINATE in Practice

Since their introduction to political scientists, NOMINATE scores have been used fairly regularly as measures of legislators' ideology. Using these scores has a number of advantages over previous approaches. First, unlike previous approaches, NOMINATE explicitly operationalizes a model of motivation. Its assumptions include (1) motivation consistent with the spatial model of voting, (2) the use of "error" information, and (3) the lack of agenda effects. Since they take the broadest set of data relating to legislator preferences--the complete set of roll-call votes from a Congress--NOMINATE scores provide a broad-based, non-time-specific measure of underlying legislator ideology. With NOMINATE, there is no need to specify in advance the coalition patterns or salient issues that structure summaries of voting behavior. With these data in hand, scholars could then arrange legislators in order of their ideology based on first-dimension NOMINATE scores and identify votes on which legislators appear to have abandoned their ideological predisposition.(2) Whether this departure is attributable to presidential intervention or something else obviously requires other controls, but such an endeavor would still be a major advance in understanding presidential-congressional relations and the nature of policy making in the federal government. Ordering legislators according to their ideology (or propensity to vote with one party or the other) could help identify a set of crucial moderates that one would expect the president to target on close votes. One could then use more refined statistical tests or perform descriptive analyses on this subset of legislators to ascertain the degree to which the president was able to influence their votes. Not only would this be more reliable than a case-by-case analysis of specific votes, it also would more accurately reflect the intuition that presidents might make long-term investments with certain moderate legislators rather than try to buy them off one vote at a time. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.