It's civics 101: Before any piece of legislation becomes law, it must first be voted on and passed by both the U.S. House and Senate. It sounds simple enough in theory, but in reality, the process is often neither simple nor straightforward.
For the last four years Larry Evans, with the help of a revolving cadre of undergraduate research collaborators, has been looking behind the generally closed doors of Congressional leadership at what he sees as one of the central investigations among political scientists today—the role that political parties and their leaders play in the legislative process. His window on this process has been the tally sheets, or "whip counts," that congressional leaders use to track the positions of members prior to major roll call votes.
"It's actually kind of difficult to measure the impact of leaders and parties on the legislative process," said Evans, Newton Family Professor of Government at William and Mary. "For instance, if you are looking at a roll-call vote, if it is an important issue more often than not you are going to have a party-line outcome with the Democrats on one side and the Republicans on the other. But the question is why?"
Evans noted that political scientists can suppose numerous reasons for the frequency of the party-line split. After all, Democrats and Republicans alike display a high degree of political cohesion, both naturally ideological or driven by interest groups or constituency concerns. Evans says another possible cause is that each party's leadership is stepping in and influencing members. The individuals who are charged with making sure there are no surprises when the votes are tallied are called "whips." In the Senate and the House, each party has a member who serves as the whip and is assisted by many deputy whips and expert staff. The whip's first job is to count votes—or rather, potential votes. Before important bills are considered on the House or Senate floor, whips often see that members of their party are individually polled, in person or by e-mail or phone.
The whip usually marks each member's response in one of five categories—"yes," "leaning yes," "no," "leaning no" or "undecided." The papers logging the responses are called whip counts and are the raw material for Evans and his students. Whips do more than just poll the members of their party. When the vote count isn't solid, whips begin the second part of their jobs, using their influence to turn some yesses into nos—or vice versa.
A public whipping
Sometimes a whip's actions are, literally, front-page news. In March, 2008, current Democrat whip James Clyburn's successful efforts to get the last few votes to pass a crucial ethics bill were very visible on the House floor. The process was televised by C-Span, reported in the Congressional Record and chronicled on page A-1 of the Washington Post.
Congressman Clyburn's high-profile whip work is the exception. Generally, Evans says, the influence of the party whips is exerted more privately. Therefore, the only way to examine how the whips do their jobs is to track changes in the whip counts. To examine patterns in whip counts and roll calls over the course of a given bill, Evans and a team of undergraduate researchers have been coding each tally into a computer database.
Evans credits the students with the bulk of the coding, a time-consuming and painstaking process. He has even conducted a summer research program for the last four years, in which students lived on campus and worked full-time on the project. Nearly 80% of the students involved in the overall project have participated in the summer program.
The students have gone through data from the whip counts of House Democrats from 1955 to 1986 and from House Republicans from 1975 to 1980 and from 1989 to 1994. Some 35 students and counting have worked on the project. No detail has been overlooked. Entries are even included for members who are absent because they are ill or out of town. To date, approximately 500,000 member responses have been coded into the database on over 800 major pieces of legislation considered before Congress since World War II.
Getting the raw data hasn't been easy. While Congressional roll call votes are public documents and recorded in the Congressional Record, whip counts are not. To get them, Evans and a number of his student researchers have traveled to archives around the country, poring over countless files, looking into the records of former Congressional whips. Evans estimates that they have photocopied more than 10,000 pages of documents.
Logan Ferree '07 traveled with Evans to visit the archives of former House Speaker Newt Gingrich at the University of West Georgia. "Although I had been interested in government when I first came to William and Mary, working on the project helped turn that interest into a passion that has resulted in a career in politics," he said in a recent email.
The researchers have also found the congressional archives to be a rich source of contextual evidence into the effectiveness of party influence wielded on House votes. This contextual information is being developed through case studies examining what factors move a whip count—for example, leadership lobbying, constituency pressure, team playing, etc. The case studies were completed in Evans' seminar classes and summer research program by student researchers such as Ricky Trotman. Trotman's initial exposure to the project was through an assignment in a freshmen seminar, American Political Development, taught by Evans.
"I did an overall analysis of a 1980s bill funding Angolan rebels. My assignment was to explain why the Democrats failed to defeat that bill. So I coded a bunch of variables like the percentage that Reagan won by in that representative's particular district, minority populations, age brackets, funding received from the party," he explained. "What it ended up coming down to is there were a large number of Hispanics in Florida and the Democrats were overwhelmingly represented in Florida at that point but they swung for the bill and it was that constituency that really put the bill over the top, which I found to be very interesting. Because for that particular whip count there had been, I think, three different polls taken. So you see a progression—maybe like an undecided, an undecided and then, ‘No, I'm not voting with you.'"
A Pickle paper
Claire Grandy '09 wrote a seminar paper on former Texas Democrat J.J. Pickle's amendment to the House's 1977 Social Security financing bill in her freshman year. "I was pretty certain (and I still suspect) that no one had ever devoted much more than 15 words to the Pickle amendment," Grandy said of the experience. "But Professor Evans walked us through the process and the research turned out to be fun—and unique."
The whip count project is funded by a large grant from the National Science Foundation and has received additional funding from the Carl Albert Congressional Research and Studies Center at the University of Oklahoma and the Roy R. Charles Center of the College of William and Mary. The project has already yielded numerous papers—many of which have been presented by the students at political science conferences—as well as two book chapters in edited volumes. A complete book is also on the horizon.
One paper written by Evans and five members of the class of 2005—Sarah Brown, Keith Devereaux, Kristen Haase, Will Marlow and Josh McHenry—received the Patrick J. Fett Award for the best paper on Congress and the Presidency in 2006 at the Mid-west Political Science Association Convention.
Since his freshman seminar, Trotman, who also participated in the summer program, has written a second paper for the project, "Parties, Preferences and Contras in the 1980s House" with Evans and co-authors Miguel Matamoros, Walter McClean and Laura Whipple. The four undergraduates and Evans presented the paper at the Western Political Science Conference in San Diego in March.
Charting the path of influence
Evans, whose research focuses on American national politics, hopes his project will not only provide scholars and others the information necessary to analyze how Congressional leadership and political parties routinely exert their influence but provide a method for tracking this influence over time.
"Because we've got this whip-count data which provides a peek at the decision-making process, you can open up the black box of member decision making in really a rich way," he said.
Followed over time, the patterns of the whip counts and final vote tallies can be telling, especially when paired with the additional contextual information being generated by Evans' project.
"The evidence we have gathered is really useful for evaluating the independent impact of the leadership, allowing us to shed light on a major scholarly debate that really has confounded most of the top scholars in the field," Evans said.
Still, he added, the papers that he and the students have written only scratch the surface in this area of research. "A long-term goal of the project is to develop and maintain one of the more valuable and highly-used research web sites about the American legislative process," he said.
In addition to influencing scholarship about Congress, the project has affected its participants.
"Most of all, through this project and my government seminar the next semester, I gained a great appreciation for good research practice and was delighted to carry over some of the data I'd worked with into coursework completing my major," said Brent Schultheis '08. "I feel more confident now that I can tackle whatever research demands I'll have in law school or the working world."
Trotman admits he was initially overwhelmed by the project, but the history and international relations major now notes he wouldn't trade the experience.
"It was definitely worthwhile thinking back on it now," he said. "Writing the paper at three or four o'clock in the morning, I wasn't thinking it was so rewarding then, but it definitely was worth it. I see a lot of value in the project and it's definitely something that can shape the future study of congressional politics."
The idea that party leaders and presidents are influential in the votes cast in Congress is not entirely a new one, but this project offers a novel way for scholars to add empirical evidence to their theories, Evans said.
"There are all kinds of stories about the techniques that leaders use to influence outcomes," he said. "But, we don't really know systematically how that occurs over a lot of bills, over a lot of years, and this evidence—gathered by all my students and me—sheds a lot of light on that."
Video feature: Selling the candidates: Evans' students create campaign ads