Berry et al.'s (1998) measures of U.S. state citizen and government ideology rely on unadjusted interest-group ratings for a state's members of Congress to infer information about (1) the ideological orientation of the electorates that selected them or (2) state legislators and the governor from the same state. Potential weaknesses in unadjusted interest-group ratings prompt the question: Are the Berry et al. measures flawed, and if so, can they be fixed by substituting alternative measures of a member's ideology? We conclude that a version of the Berry et al. state government ideology indicator relying on NOMINATE common space scores is marginally superior to the extant version. In contrast, we reaffirm the validity of the original state citizen ideology indicator and find that versions based on NOMINATE common space scores and adjusted ADA and COPE scores introduced by Groseclose, Levitt, and Snyder (1999) are weaker.

We construct measures of U.S. state partisan identification, self-reported ideology, and policy mood using data from the 2000 and 2004 National Annenberg Election Surveys (NAES) and the 2006 Cooperative Congressional Election Study (CCES). These measures improve on existing methods for estimating state-level preferences because the surveys provide larger state samples without pooling across years. After detailing our methods for constructing the measures, we assess their validity through comparisons with measures already in use by scholars of state politics. We find that our measures correlate strongly with those created by Erikson, Wright, and McIver (1993) and Berry et al. (1998) and with measures from state-level polls. We conclude that our measures can be useful to research in state politics.

The introduction of McDonald and Popkin's (2001) measure of the voting eligible turnout rate represents an important advancement in the literature on voter turnout in the United States, most of which was based on studies of turnout among the voting age population. Among other things, their measure called into question many of the findings of research on the decline in turnout during the 1960s, 1970s, and 1980s. In this article, we investigate how the measurement of voter turnout affects the interpretation of the sources of turnout in the states. Focusing on voting-age and voting-eligible turnout in presidential elections from 1980 to 2004 as separate dependent variables, we estimate models of state-level turnout and compare the results for each of the measures. Our findings suggest that measurement does matter in state-level voter turnout models and that some inferences drawn from studies of turnout among the voting age population are not sustained when substituting voting eligible turnout.

Many hypotheses in U.S. state politics research are multi-level, positing that state-level variables affect individual-level behavior. Unadjusted standard errors for state-level variables are too small, leading to overconfidence and possible false rejection of null hypotheses. Primo, Jacobsmeier, and Milyo (2007) explore this problem in their reanalysis of Wolfinger, Highton, and Mullin's (2005) data on the effects of post-registration laws on voter turnout. Primo et al. advocate the use of clustered standard errors to solve the overconfidence problem, but we offer an alternative solution: randomization tests. Randomization tests are non-parametric tests that do not rely on comparisons to theoretical test statistic distributions. Instead, they use distributions tailored to the data, created by randomly scrambling the data many times to simulate what would be observed under the null hypothesis. Unlike with clustering, with the randomization test, U.S. state-level reforms generally fail to be significant both as additive effects and as interactions with individual characteristics.

By measuring U.S. term limits dichotomously, investigators ignore the vast differences among laws limiting state legislative service. Furthermore, this measurement problem increases the risk of false negatives and confounds the effects of term limits with those of the citizen initiative. To address this, I propose two sets of continuous measures of term-limitedness. The first set compares mandated turnover after term limits to turnover in the 1980s, the decade before term limits began sweeping elected officials from office. A second set adjusts the first set to reflect the potential for legislators to cycle repeatedly between legislative chambers when only their consecutive years of service are limited. These continuous measures outperformed a dichotomous designation of term limits in two tests, suggesting that the proposed measures can reduce the risk of false negatives about term limits in U.S. multi-state research and that they are more robust in the face of confounding effects from the citizen initiative.