by David C.W. Parker
In 2011, Montana’s lone Congressman, Republican Denny Rehberg announced that he would run for the Senate seat held by Democratic freshman Jon Tester. The race was one of the most competitive Senate contests during the 2102 cycle, with nearly every public poll conducted over the next 20 months showing the race tied statistically. After more than $50 million spent on television, radio, mail, and GOTV by the parties, candidates, and nearly 30 outside groups, Tester won reelection by less than four points.
I was fortunate to have a front row seat for the campaign. Both Congressman Rehberg and Senator Tester allowed me unprecedented access during the race and, after the campaign, to their staffs. I documented in my forthcoming book with CQ Press, Battle for the Big Sky: Representation and the Politics of Place in the Race for the U.S. Senate, the electoral consequences of the representational relationships both candidates had established with Montanans. Much like Richard Fenno’s pioneering work, I soaked and poked my way through the campaign while documenting the pre-political and early political careers of the candidates. Unsurprisingly, I conclude that campaigns can affect the outcome of elections.
As I reflect about the many lessons I learned on the campaign trail, one stands above the rest. As many of you watch competitive Senate races this cycle, you’ve probably heard that Democratic Senate candidates vote with Barack Obama a lot. Congressman Cory Gardner in Colorado, for example, tells Coloradoans that Senator Mark Udall votes with Barack Obama 99 percent of the time according to his 2013 CQ Presidential Support score. Here in Montana in 2012, ads sponsored by Rehberg, the Republican Party, and outside groups cited ad nauseam the claim that Senator Tester voted with Obama 95 percent of the time.
One of my clearest memories during the campaign was an exchange I had with Aaron Murphy, Tester’s communications director. Murphy claimed that the Presidential Support Score calculated by CQ was some “random number” calculated by some Washington insiders. I called Aaron to dispute this. It is not some “random number”, but a metric that is carefully calculated according to precise rules and it is widely employed by political scientists studying Congress. It has face validity and provides a good thumbnail sketch of how supportive members of Congress are (or not) of an administration. Aaron pushed back, saying that the number does not, by any means, tell the whole story of Tester’s record. In fact, Tester—as Montana’s Senator—had developed a solid record of independence that is not reflected by the 95 percent figure.
It turns out that Aaron was right. Not that the number is random, but that it is a flawed metric.
The presidential support score, as calculated, does not tell us as much as we might think about a representative’s relationship to a presidential administration. And, if I might be so bold, political scientists are doing a great disservice to the study of politics by oversimplifying the legislative process by repairing to mindless empirical reductionism. Make no mistake: I am not the naysaying baseball scout Grady Fuson from the movie “Moneyball.” Quantification has allowed political science to develop tremendous insight into the practice of politics. But, at the same time, quantification run amok has unnecessarily narrowed the research questions we ask, the research agendas we pursue, and creates a distance from politicians and politics that hinders the construction of new theory, new observation, and new metrics.
Consider the Presidential Support Score. It records the votes of members of Congress on issues on which the administration sees fit to take a stance publically. The score includes some votes that are essentially non-controversial—and this is particularly true in the case of the Senate, where many votes to appoint executive branch officials and judges receive widespread bipartisan support. As I note in Battle for the Big Sky, Tester’s rating of 95 percent drops to 85 percent once these noncontroversial votes are removed from the mix using a procedure established by Bond, Fleisher, and Northrup (1992).
But, more distressing, is the fixation political science has with this blunt measure of such a complex phenomenon. As political scientists, we accept as axiomatic that roll call voting only represents a portion of the jobs that members of Congress do and only a slice of position-taking type activities. David Mayhew (2000) cleverly reminds of this by developing a richer metric of the ways in which members of Congress engage in the public sphere. Mayhew canvasses the historical record for examples of the actions that members of Congress take that receive the attention of historians. Only half of the actions taken by members of Congress in the public sphere are position-taking on roll call votes. Emphasizing roll call votes—which are easily obtained and scaled—puts too many of our empirical eggs in one basket and biases our view of what legislators do.
To return to the presidential support score, there are a host of other ways that members of Congress can and do demonstrate their relationship to an administration. Take, just as one example, the communications members of Congress send to executive branch officials. Members of Congress often write to agencies before, during, and after the rule making process to protest certain administrative actions. In 2012, when the Department of Labor attempted to apply workplace safety rules to family farms, the Obama administration heard loud protests from Montana’s congressional delegation. Tester alone introduced legislation to stop the rules from going into effect, sent letters to the Secretary of Labor Hilda Solis protesting the rules, and took to the Senate floor to remonstrate against the Department of Labor’s regulations. None of these position-taking actions were represented by a roll call vote in opposition to the administration. I’ll leave it to the reader to decide if this reflects 95 percent support of Barack Obama.
And even roll call votes themselves are increasingly misleading as omnibus legislation and resolutions dominate a gridlocked legislature. In mid-September, the House and Senate both approved an appropriations measure that would continue to fund government after October 1, 2014. That same resolution contained language and funds allowing the administration to train and aid Syrian rebels, “$88 million in additional funding to combat the Ebola epidemic, $64 million for the Department of Veterans Affairs and $6 million for the Commodity Supplemental Food Program, which helps low-income people obtain nutritious food” (Shabad and Cox 2014). The administration supported the bill, but did a vote for the funding of government signify support for the administration? Did it signify the desire merely to avoid a shut down? Did it signify support for the president’s request to authorize air strikes? Was it a vote to fight Ebola or help poor people in need of food? That vote carried multiple meanings, but it will probably be scored by CQ as supporting the administration even though a vote for it or against it could explained variously depending upon where one sits.
The outlook is not all dim for political science and the study of representation. Justin Grimmer’s (2013, 2014) recent work shows how the tools of textual analysis can illuminate how members of Congress chose to communicate and connect with constituents. Greg Koger’s (2010) wonderful book on the filibuster emphasizes sequence, context, and timing whilst taking aim squarely at the wrongheaded assumption of single-peaked preferences. Neither follow politicians around, but both demonstrate both a complex understanding of politics combined with sophisticated quantitative analyses.
But we don’t have to rely upon Stata and R alone to make insights into the political process. Spending time with politicians reminds us that representation and legislating are complex concepts difficult to nail down precisely with just one or two measures. The more exposure we get as political scientists to how politicians do their jobs, the more we see and the better the theories we develop. As I write in the final chapter of my book, political scientists just don’t “do” Fenno anymore. That’s a shame because Richard Fenno’s soaking and poking did not exhaust the sights for political science to take in. There is still much to see and learn. We ought to start doing it—on the campaign trail, in Washington, and through the use of congressional archives where we can observe the second face of power. The end result makes our research more accessible, more relevant, and ultimately, theoretically richer.
Dr. Parker is an associate professor of political science at Montana State University. He is the author of The Power of Money in Congressional Campaigns, 1880-2006 (Norman: University of Oklahoma Press), as well as articles on the consequences of divided government and how members of Congress build reputations with their constituents. His article, “Making a Good Impression: Resource Allocations, Home Styles, and Washington Work,” won the 2010 Alan Rosenthal Award from the American Political Science Association. His co-edited volume on archival research methodology, Doing Archival Work in Political Science, was published by Cambria press.