A note on social capital and network content

As a guide to selecting name generators for social capital research, I use network data on a probability sample of heterogeneous senior managers to describe how they sort relations into kinds, and how the kinds vary in contributing to social capital. Managers sort relations on two dimensions of strength - intimacy (especially close versus distant) versus activity (frequent contact with new acquaintances versus rare contact with old friends) - and with respect to two contents - personal discussion (confiding and socializing relations) versus corporate authority (the formal authority of the boss and informal authority of essential buy-in). Comparing name generators for their construct validity as indicators of social capital, I compute network constraint from different kinds of relations, and correlate constraint with early promotion. The correlation is strong for the network of personal relations, zero for the network of authority relations, and strongest for personal and authority relations together. I close with research design recommendations for selecting name generators.     

Relation contents in multiple networks

Distinctions among kinds of relations (friendship, advice, intimacy, and so on) are typically ad hoc in empirical research. These ad hoc distinctions among relation contents can be expected to increase the likelihood of equivocal research conclusions. We develop three ideas indicating how standard, well-known, network models of relationship form can be used to clarify relationship content. (a) We begin with an idea for recovering the semantic context in which a relation content occurs. This context is cast as a network of tendencies for contents to be confused for one another and the form of this network—dissected with network models of relation form—holds insights into the ways in which relation contents are understood in a study population. (b) The network concept of structural equivalence is used to define content domains composed of specific relation contents that are substitutable for one another in described relationships. (c) The network concept of network prominence is used to define the ambiguity of contents in described relationships. The proposed perspective is analogous to a linguistic componential analysis of relation content.     

Comparative power structures in American communities

This is an argument for the plausibility of comparative network analyses of power in American cities. In terms of structural analyses of community power and decision-making, traditional methods of obtaining network data have fostered and enervating gap between the substantive richness of case studies and the generalizability of comparative studies. The purpose here is to describe a practical method of gathering ersatz network data from interviews with a small number of strategically chosen informants so as to conduct comparative network analyses of community power. The ersatz network element z_ij is the extent to which two types of actors I and J are perceived by informants as having simultaneous control over community issues. Using data from interviews with informants in 51 American cities, the ersatz network data are shown to be efficient, reliable and substantively valid despite their low cost. The data are efficient in the sense that the proportion of sampling variance in an estimated network element declines exponentially with increase in the element's magnitude. The data are reliable in the sense that estimated structures are stable over issue and informant bias per se. The data are valid in the sense that they are consistent with previous community research; types of ersatz power structures are located in the 51 cities (a machine structure, an urban structure, a faction structure, and a class of reform government structures) and these types occur in cities with characteristics known to be associated with each type of power structure. I conclude with a summary and comments regarding future research; the selection of informants and the use of a computer program in which easily available characteristics of a community are the basis for predicting the type of power structure it most likely exhibited during the late 1960s.     

Bayesian Binomial Regression: Predicting Survival at a Trauma Center

Standard methods for analyzing binomial regression data rely on asymptotic inferences. Bayesian methods can be performed using simple computations, and they apply for any sample size. We provide a relatively complete discussion of Bayesian inferences for binomial regression with emphasis on inferences for the probability of “success.” Furthermore, we illustrate diagnostic tools, perform model selection among nonnested models, and examine the sensitivity of the Bayesian methods.