Elected politicians can manage government agencies using administrative procedures that either enhance agencies' political responsiveness or their autonomy. Rational politicians will delegate autonomy when the benefits of agency decision-making are high, such as when the policy problem is technically complex. This article tests this hypothesis by investigating how state politicians design the institutional features of state clean air agencies and how these agencies perform under different administrative procedure regimes. I find that when a state's air pollution problem is complex, politicians eschew administrative procedures that enhance political control, and these agencies respond by choosing policies from a broader range of options.

Can a school curriculum influence the political socialization of students? And if so, can such socialization “trickle up” to influence the political socialization of these students' parents? We examine the effects of Kids Voting USA in Kansas to answer these questions. By the 1996 presidential election, Kids Voting was implemented in several Kansas counties. Consequently, we are able to compare turnout—and change in turnout across different elections—in such counties with those that did not adopt the program. After controlling for a variety of county-level characteristics that are likely to affect turnout, regression analyses indicate a positive and significant contribution of Kids Voting to county turnout, and change in turnout across elections. We address the practical implications of this study and emphasize its theoretical contribution to theories of political socialization.

Retail sales on the Internet are growing at a rapid pace. Some states have tapped into this potential revenue source with Internet sales taxes, while others have not. What are the factors that lead states to make these policy decisions? Using a 50-state comparative research design, we find that interest groups, as much as the partisan and economic factors emphasized in previous research on state tax policy, are correlated with adoptions of these state tax innovations. Furthermore, this interest group influence is not limited to a single industry, but comes from self-interested groups representing both sides of the issue, including newly emerging technology interests.

Can minority representation influence policy outputs when decisions are largely governed by formal and informal rules? This article addresses this question using detailed data on school district representation and budgetary allotments in Texas, assessing the link between Hispanic school board representation and budgetary support for bilingual education. My findings suggest that descriptive representation does affect the budgetary process. My analyses also suggest that modeling representation as a role rather than as an input provides greater insight into the representational process.

This article examines candidates' perceptions of political party services provided in state legislative campaigns using survey data from seven states in 1994. I find that candidates generally do not feel that party assistance is very helpful, although multivariate analyses show that this varies systematically with state context and district-level competition. The influence of campaign financing on candidate perceptions is particularly important. In districts where candidates spend similar amounts on their campaigns, candidates believe parties provide greater assistance. My findings suggest that parties target campaign services to candidates in the same way that they target financial contributions, focusing on districts where these resources are most likely to affect the election outcome.

Elite interviews have long been a staple of state politics research. Improved information technology facilitates the use of elite interviews, but also underscores the need for attending to their design, operationalization, and analysis. This essay provides a framework for developing, conducting, and interpreting elite interviews, and suggests means to enhance the validity and reliability of findings by considering instrumentation, sampling, data collection, and transcript analysis.