We combine two foci of interest with respect to community identification and node centrality and create a novel metric termed “leadership insularity.” By determining the most highly connected nodes within each community of a network, we designate the ‘community leaders’ within the graph. In doing this, we have the basis for a novel metric that examines how connected, or disconnected, the leaders are to each other. This measure has a number of appealing measurement properties and provides a new way of understanding how network structure can affect its dynamics, especially information flow. We explore leadership insularity in a variety of networks.

This paper presents a new measure of centrality, scalar products centrality that is appropriate for dense networks in which link strength is measured with real numbers rather than by a simple dichotomy. Scalar products centrality may be defined by a node’s distance from the center of the set of measured relations that compose a network. Formulas for its calculation based upon the centroid scalar products matrix from classical multidimensional scaling are presented. Two examples are provided and the measure is compared with standard measures of centrality (degree, eigenvector centrality, betweenness and closeness) to demonstrate its validity. As expected, the measure is strongly related to the degree and eigenvector measures and less so with betweenness and closeness.

Centrality measures are often used to describe influential nodes in a network. When these measures are highly correlated they may be redundant and when they are uncorrelated they provide unique insight into the network. I propose a network simulation approach that creates networks with varying degrees of Erdos-Renyi randomness and Albert-Barabasi scale-freeness. Using this simulation approach I conduct 10 replications of a full factorial experimental design with varying levels of density and randomness versus scale-freeness. The effects of topology and density on the correlations of degree, betweenness, closeness, and eigenvector centrality are investigated. I find that not only does density and topology affect the correlation of centrality measures, but there exist many interaction effects as well. In general, networks with high Erdos-Renyi randomness tend to have higher levels of correlation between centrality measures than networks with Albert-Barabasi scale-freeness.

This paper compares two models of interstate influence: the proximity model (which posits that states are more influenced by nearby states than farther ones) and the opinion-leader model (which hypothesizes that some states are “regional leaders” which exert a disproportionate influence on all other states). These models are compared for two outcomes, reproductive health policies and general liberalism, using regression models enhanced by network analysis techniques. Data on the 50 states were collected. Dependent variables included policies and spending on reproductive health as well as a broader range of policies designed to measure general liberalism. Independent variables included historical and geographic conditions, socioeconomic factors, political behavior, governmental institutions, and the behavior of elites. Results indicate that a state’s policies on reproductive health (excluding abortion) appear to be a function of both the socio-economics of the state as well as the reproductive health policy of its regional leaders. General liberalism towards reproductive health, in contrast, is largely explained by per capita income and percentage of the state population that is fundamentalist, as well as the liberalism of a state’s geographic neighbors. The importance of acknowledging network effects in state analyses, and of progressive states’ leadership in advancing reproductive health, is underscored.

This study considers the extent to which work-related social ties function as a source of social support in Russian workers’ personal networks. The topic is important since, in the case of unemployment or retirement, personal networks are central for the well-being and coping of Russians. In order to illustrate the nature of the Russian case, an explicit comparison between Russian and Finnish workers’ personal networks is carried out. The results are in line with previous findings concerning the workplace as a source of social support in China.

Interactions, exchanges of ideas and cooperation among scholars are important factors for the advancement of scientific knowledge. Conferences represent one of the most suitable occasions to further scientific interactions, stimulated through the contributions presented either by a single researcher or by a group of authors. Using the books of abstracts from four recent Italian conferences on population studies (Giornate di Studio sulla Popolazione, GSP, 1999, 2001, 2003, and 2005), this research provides an empirical analysis of the collaboration patterns observed among the authors of the papers presented. We followed a social network perspective, in order to find out the determinants of scientific cooperation in the field of Italian demographic studies. The factors playing a major role in determining the actors’ relationships seem to be related to gender and to the proximity of universities or conference seats. Although a high number of participants are represented by isolated nodes, the most common way of collaborating is a dyadic relationship. The larger collaborations are due mostly to the presence of a small number of leading authors that manage a large number of papers. Productivity and the popularity of leading authors are attributed to their senior positions in research groups or their technical and statistical skills. It is difficult to measure such aspects with an analysis approach that is different from network analysis.