Democracy's Voices: Social Ties and the Quality of Public Life in Spain (review)

Robert M. Fishman in his book, Democracy's Voices: Social Ties and the Quality of Public Life in Spain, traces an interesting picture of labor politics in Spain as he explores the relationship between social network ties and the quality of democratic life. Fishman focuses attention squarely on his primary independent variable – the ties between workers and intellectuals – wherein intellectuals are identified as "authors who write for an educated or general audience." His primary claim is that worker-intellectual ties shape the quality of democratic life because they point leaders toward one of two distinct styles of labor politics – defensive localism or global discursive horizons. Defensive localism implies that leaders focus on local concerns by expressing decidedly local objectives, while leaders adopting global discursive horizons attempt to connect specific, local concerns to broader and more general, extra local objectives. Fishman compellingly elaborates on the differences between these two political styles with a rich comparison of the struggles of labor in numerous communities.

Fishman's research strategy involves analyses of survey data from a snowball sample of leaders in 49 industrial, working-class communities in Spain, supplemented with qualitative interviews. Fishman treats readers to a fascinating discussion of the character of work-related politics in Spain, as well as an analysis of the Socialist and post-communist subcultures through which the influence of worker-intellectual ties operate. His empirical analyses also highlight the enduring effects for those participating in the opposition movement during the Franco period – a finding that resonates with McAdam's work on Freedom Summer participants. Overall, Fishman assembles a persuasive case for the important influence that worker-intellectual ties have on the democratic process in Spain.

Of course, any attempt to link social ties to democratic outcomes will eventually lead social scientists to thoughts of social capital. Fishman tackles social capital with surprising and convincing ferocity in chapter four. He argues that the concept of social capital is problematic because it is multidimensional, it masks important distinctions between crucial indicators, and its usage implies the misguided assumption that social relations represent instrumental investments. While his argument will no doubt ruffle a few feathers, his forceful discussion underscores the importance of considering carefully concepts that become vogue.

The multivariate analyses presented in chapter six were a disappointment. The author uses logistic regression to predict the dichotomous outcome of whether local leaders articulate globalizing discursive horizons or defensive localism. In addition to including variables that were considered in bivariate analyses in previous chapters, the author introduces a series of control variables. Unfortunately, these control variables produce confusion for readers because their measurement is not carefully discussed. In particular, the variables that indicate three different institutional positions of the respondents in leadership roles initially appear to be a set of dichotomous dummy variables. Including a set of dummy variables in a logistic regression model is a straightforward procedure, with the interpretation of the dummy coefficients involving a comparison to an omitted category. In contrast to the typical presentation of a single model with a focus on comparisons to the omitted category, the author reports two models, each with a different omitted institutional position variable, "In order to assess the causal impact of all three institutional locations." (p.134) The coefficients presented in these models, as well as the author's interpretation of these estimates, led me to believe that these institutional variables do not constitute a set of dummy variables as typically understood. In one table, the reported N differs between the two models. A more detailed description of the institutional position variables, as well as the other control variables, would be of great interest to readers. Additionally, while the bivariate analyses highlight the usefulness of measuring the strength of ties as a three-category variable (no ties, simple ties and multiple ties), the multivariate models appear to treat tie strength as a continuous variable. I would like to see an explicit examination of how...