The forthcoming article "Congressional Representation: Accountability from the Constituent's Perspective" by Stephen Ansolabehere and Shiro Kuriwaki is summarized by the author(s) below.  

Does voting against your constituents on a bill jeopardize a representative's re-election prospects? Recent party polarization seems to cast some doubt. Perhaps party attachments have become so entrenched that a co-partisan voter will support their member's re-election regardless of how the representative votes on the issues. And how much do voters know about member's specific issue stances in the first place? 

Our AJPS article is a novel 14-year study conducted as part of the Cooperative Congressional Election Study (CCES). We use a mix of observational analysis and two survey experiments.   The innovation in this study was to ask citizens their preferences about issues that Congress voted on and how they thought their members voted on those issues. 

Three findings stand out: 

  1. At the individual level, constituents hold legislators accountable for their roll call votes.  A representative's roll call vote in agreement with a constituent's preference leads to an increase in that constituent's approval for the representative by 11 percentage points.  This number accounts for the fact that 4 out of 10 constituents know how their representative voted on an average roll call (the other 4 out of 10 are not sure; only 2 out of 10 are incorrect). Among respondents who correctly update their perceptions due to a rollcall vote, the effect is about three times larger. 
  1. This association is substantively meaningful. An effect of 11 percentage points on approval is comparable to the effect of having a copartisan representative. When it comes to vote choice rather than an approval of an incumbent, partisanship weighs larger than specific votes (17 points) but agreement on issues is still important (10 points). 
  1. But at the congressional district level rather than an individual constituent's level, a more muted picture of accountability arises.  Districts are mixed in issue preferences and party identification. On any vote, there are many people in a typical district who support one side and many who support the other. In the average congressional district, the two parties split their support 60 – 40, rather than clump into entirely Democratic and entirely Republican districts. As a result, most individual accountability cancels out and the shift of vote share on elections is about 2 percentage points. 

This last finding is consistent with studies of the responsiveness of election results to shifts in legislators' roll call voting patterns.  Those studies, using aggregate data, might imply that individual voters are not strongly moved by issues. Our study, thus, resolves a puzzle – how is it that individual citizens care deeply about issues but aggregate correlations between election results and legislative voting are modest?  Voters are attentive and issue-oriented, as well as partisan, in their behavior.  The modest effect of issues (and partisanship) at the district-level reflects the diversity of opinion in the typical congressional district in the United States.  If legislators were to change their votes on key pieces of legislation, they will lose support among some constituents but gain among others.  Aggregate representation reflects the share of people who support the legislators' decisions net those who oppose those decisions.  Our study, then, provides the micro-foundations of aggregate representation.    

Our analysis reveals both the power of democratic accountability and its blind spots. Obfuscation can dampen accountability, as we see in our data in the 2014-2015 Congress where Congressional stalemate led to few key votes getting a vote. What this means for Congressional policymaking and how opinions get aggregated into districts is the subject of our ongoing work. There we comprehensively studies how issue preferences are aggregated and acted upon by Congressional representatives, the single member district (FPTP) electoral system, and the Congressional agenda. 

About the Author(s): Stephen Ansolabehere is Frank G. Thomson Professor, Department of Government at Harvard University and Shiro Kuriwaki is PhD candidate, Department of Government at Harvard University. Their research "Congressional Representation: Accountability from the Constituent's Perspective" is now available in Early View and will appear in a forthcoming issue of the American Journal of Political Science.