for research assistance. 778 778 778 778 AMERICAN SOCIOLOGICAL REVIEW AMERICAN SOCIOLOGICAL REVIEW AMERICAN SOCIOLOGICAL REVIEW AMERICAN SOCIOLOGICAL REVIEW AMERICAN SOCIOLOGICAL REVIEW 1 To be sure, many of these countries have incomplete or "façade" democracies without fully competitive elections (Markoff 1996, chap. 5). Even within the most democratic countries, barriers to participation inevitably persist (e.g., registration requirements, barriers faced by disabled voters, difficulties accessing polling places, especially when elections are held on working days). Every country excludes noncitizen immigrants from voting in national elections. 2 We thank Joe Levinson at the Prison Reform Trust, and Femke van der Meulen at the International Centre for Prison Studies, both in London, for making the results of their international survey of felon voting rights in Europe available to us.

Convicted felons face both legal and informal barriers to becoming productive citizens at work, responsible citizens in family life, and active citizens in their communities. As criminal punishment has increased in the United States, collateral sanctions such as voting restrictions have taken on new meaning. The authors place such restrictions in comparative context and consider their effects on civil liberties, democratic institutions, and civic life more generally. Based on demographic life tables, the authors estimate that approximately 4 million former prisoners and 11.7 million former felons live and work among us every day. The authors describe historical changes in these groups; their effects on social institutions; and the extent to which they constitute a caste, class, or status group within American society. The authors conclude by discussing how reintegrative criminal justice practices might strengthen democracy while preserving, and perhaps enhancing, public safety.

Criminal offenders in the United States typically forfeit voting rights as a collateral consequence of their felony convictions. This article analyzes the origins and development of these state felon disenfranchisement provisions. Because these laws tend to dilute the voting strength of racial minorities, we build on theories of group threat to test whether racial threat influenced their passage. Many felon voting bans were passed in the late 1860s and 1870s, when implementation of the Fifteenth Amendment and its extension of voting rights to African-Americans were ardently contested. We find that large nonwhite prison populations increase the odds of passing restrictive laws, and, further, that prison and state racial composition may be linked to the adoption of reenfranchisement reforms. These findings are important for understanding restrictions on the civil rights of citizens convicted of crime and, more generally, the role of racial conflict in American political development. Punishment for felony-level crimes in the United States generally carries collateral consequences, including temporary or permanent voting restric

Do mass policy preferences influence the policy output of welfare states in developed democracies? This is an important issue for welfare state theory and research, and this article presents an analysis that builds from analytical innovations developed in the emerging literature on linkages between mass opinion and public policy. The authors analyze a new dataset combining a measure of social policy preferences with data on welfare state spending, alongside controls for established causal factors behind social policy-making. The analysis provides evidence that policy preferences exert a significant influence over welfare state output. Guided also by statistical tests for endogeneity, the authors find that cross-national differences in the level of policy preferences help to account for a portion of the differences among social, Christian, and liberal welfare state regimes. The results have implications for developing fruitful connections between welfare state scholarship, comparative opinion research, and recent opinion/policy studies.

We present evidence of a historic realignment in the relationship between class and voting behavior in U.S. presidential elections in the postwar period. We take advantage of recent advances in class analysis and statistical methodology to introduce a distinction between "traditional" class voting and "total" class voting. Neither shows a decline in the postwar era. The realignment occurred since 1968, as professionals and nonmanagerial whitecollar workers movedfrom voting for Republicans to supporting Democratic presidential candidates. Stronger'support for Republicans among the selfemployed and among managers has more than offset the shift of professionals and nonmanagerial white-collar workers to the Democrats. Skilled bluecollar workers have become volatile, moving away from their historic support for the Democratic Party withoutfirmly attaching themselves to the Republican Party. Significant class differences in voter turnout also contribute to the total association between class and voting outcomes. Scholars and political activists have long debated the role of class divisions in U.S. politics. At mid-century, research on the "democratic class struggle" (Anderson and Davidson 1943) was a central concern of the sociology of politics. The seminal studies of political behavior in the 1950s featured detailed analyses of the organizational and sub