Research Articles

In a nationwide survey Carey, Niemi, Powell, and Moncrief (2006) found that termlimited state legislators feel less constrained by their constituencies. I use direct measures of legislative activity to examine how this "Burkean shift" in attitudes is manifested in roll-call behavior. With a new dataset consisting of all competitive state legislative roll calls for the 1999–2000 sessions and a new measure of district constituency preferences, I examine three hypotheses: that term-limited legislators are less representative of their constituents, are more polarized, and participate less in roll-call voting. I find no evidence that term-limited legislators are any less representative, and no differences in levels of party polarization appear associated with the term limits reform. I find that the impact of term limits on roll-call voting is manifested in decreased legislative effort, but this effect only appears in the more demanding legislatures. The results are consistent with the sorting model in which elections are reasonably efficient at selecting leaders whose preferences align with those of their districts, but the prospect of re-election has little role in achieving representation of constituents' ideological preferences.blic debate.