The two books under review return to the very origins of the modern empirical study of electoral behavior, the Columbia School of electoral research. Among the first studies of electoral behavior utilizing survey data were those undertaken in the 1940s by a group of scholars led by Paul F. Lazarsfeld at Columbia University. One of their main findings was that the flow of politically relevant information through varying channels and processes, whereby voters receive information about partisan politics during the course of electoral campaigns, were important determinants of the voting decisions.

One of these channels was identified as direct transmission of political information through face-to-face contact within personal networks. In Partisan Families, Alan Zuckerman, Josip Dasovic, and Jennifer Fitzgerald focus on the most intimate of social units, the family. The data sources are the (West) German Socio-Economic Panel Study (GSOEP), 1985–2001, and the British Household Panel Study (BHPS), 1991–2001. The samples in these surveys are large.

The data sets contain information about party identification, political interest, membership in voluntary organization, social contacts, and trade union and other social memberships that reveals social ties beyond the household. It is also possible to control for the class and religious cleavages in the analyses of political influences within the families. The research design is based on interviews of everyone present in the household over the age of 15, rather than relying on the reports of only one member, such as the head of household. These household surveys provide information about the most intimate of social units, but they offer few details on other forms of social life. There are, for example, no direct questions on other social networks or discussion groups and no information about the frequency of political discussion within the families.

The empirical analysis focuses on the two dominant parties in the two countries. In most of this analysis, the dependent variables are partisan support (whether the respondent supports a party or not) and the choice of parties. The focus is on the nature of partisanship in the two countries; the influence of sociostructural variables and the partisan support and choice of the other family members on the support and choice of the head of household; how husbands and wives affect each other's partisan support and choice of party; and how parents and children influence each other within families. A main finding concerns bounded partisans. This notion takes issue with the idea of a stable party identification from the Michigan School of electoral behavior: Partisanship is far from stable. Over time, most people effectively turn their backs on one of the major parties by never choosing to support it, but they switch between support of the other large party and support of no party.

Partisan choice and the constancy of partisanship over time of the head of household are directly and strongly affected by the partisan preferences of others in the household. These are stronger predictors than social cleavages and other social background variables. Husbands affect wives and wives affect husbands, but the strength of this dyadic relationship varies across countries. In Germany the male partner has a much stronger influence on his spouse, but in Britain this relationship is reversed, though the asymmetry is less marked than in Germany. The analysis of the influence between parents and children concludes that wife/mother stands at the center of family partisan politics. She always influences the husband's and children's partisan choices.

Partisan Families is in many respects an impressive book. The authors exploit the great potential of panel data for examining the reciprocal influences within families with advanced statistical methods. However, the analysis is weakened by the exclusion of several variables due to lack of indicators: Do the family members share the same value orientations and issue positions when they support the same party? What role does political discussion play in influencing family members? And what is the relative impact of personal contact with family members and others on party choice?

Democracy, Intermediation, and Voting on Four Continents presents results of comparative analyses of electoral behavior in 13 countries. It is the product of the so-called Comparative National Election Project (CNEP), which involves collaboration among teams of scholars on different continents and is based on a common core of items included in national election surveys conducted in old and new democracies in Europe, North and South America, and Asia.

The scope of the book is considerably broader than that of Partisan Families for two reasons: 1) It includes analyses from both established democracies and countries that were democratized in the course of the “third wave” of political reforms. 2) It focuses on all three channels of flow of politically relevant information in election campaigns that were central to the Columbian approach. These channels—called “political intermediations” in the book—are face-to-face contact within personal networks, the flow of messages through mass communication media, and indirect flow of political communication through membership in secondary associations. The core of the book for examining the political intermediation processes are the chapters focusing on each of these channels, as well as another chapter that examines the impact of intermediations on voting.

The chapter on mass media—by Rüdiger Schmitt-Beck and Katrin Voltmer—examines the consequences of mass media for six orientations toward the political system in the six new democracies from the third wave of democratization. The main findings are that the media do matter and that the impact of media is beneficial, not detrimental, to the consolidation of new democracies. It is primarily the information-rich media, especially high-quality and middle-stratum newspapers, that make the most positive contribution to the political orientations and, in turn, the quality of democracy.

The chapter on intermediation through secondary organization, by Paolo Bellucci, Marco Maraffi, and Paolo Segatti, analyzes 10 countries, both established democracies and third-wave democracies, and focuses on the pattern of membership in voluntary associations, respondents' reported political contacts at the time of elections by associations to which they belong, the extent to which the political information disseminated by organizations is explicitly supportive of a party or candidate, and the impact of various dimensions of organizational membership on turnout. The most valuable aspect of the chapter is the comparative analysis of various dimensions of organizational membership and the analysis of the web of the associations' political contacts, as well as how important cross-national patterns are explained by macro-level variables.

In a rich and interesting chapter, Bradley Richardson and Paul Allen Beck examine the political messages received by voters from personal discussants and the mass media in 12 countries. This chapter most clearly addresses the same topics as in Partisan Families. However, the chapter addresses citizens' exposure to political information not only from spouses but also from other personal discussants, newspapers, and television. It also examines voters' perceptions of the partisan bias of the political messages they receive from each intermediary source.

It is difficult to compare the findings in Richardson and Beck's chapter with those in Partisan Families, although both Britain and West Germany are included. However, two of their findings can place the Partisan Families focus on the influence within the families in a broader perspective: 1) Perceived partisan bias from newspapers and other discussants is greater than from spouses in both countries. 2) Personal discussions are usually more likely than the media to have congruent partisan biases with the respondent, but this is not unique for discussions with the spouses. The perceived partisan agreement with others with whom the respondent discusses politics is also just as frequent in West Germany and Britain. A main finding characterizing most countries included in the study is that of a model of pluralistic voter information contexts that lack strong and homogeneous signals.

In a long and informative chapter, Pedro Magalhães examines whether the informational biases to which voters are exposed during electoral campaigns influence their voting choices. The perceived biases for the governing versus the opposition parties are measured for four intermediaries, and the chapter analyzes the impact of these variables to explain voting behavior. A main finding is the extent to which the perceived biases in interpersonal discussion networks contribute to an explanation of the vote. Interpersonal discussion biases have a statistically significant impact in all 11 countries included in this analysis, but political discussion with spouses is excluded from the measure of interpersonal discussion. The findings, therefore, cannot be compared with those in Partisan Families.

In addition to these chapters that directly examine the political intermediation processes inspired by the Columbia School approach, the book contains two chapters on attitudes toward democracy based on data from seven third-wave democracies and a very long chapter on value orientations and party choice. These are valuable contributions within their field, but it could be questioned how well they fit into the book.

Democracy, Intermediation, and Voting stands out as an important book. Its main strength is the wealth of empirical information about different political intermediaries and their impact on voting and on political cultural variables, as well as the careful control for all types of background variables in order to examine the causal impact of the various intermediaries.

The two books represent a renewal of a political sociological tradition that has long been neglected. In contrast to the original Columbia studies, which relied on subsamples in the United States, both books use comparative representative national surveys to analyze political intermediaries. Further, both reflect the enormous advance in statistical methods that has taken place during the decades since the original studies.