Search

In the fall of 1994, following a major marketing campaign, this book appeared on virtually all of the best seller lists in the United States. The engaged public response to [...]

1. INTRODUCTION Let me begin by noting that I found the article by Such-man and Jordan stimulating and exciting. The authors are to be congratulated for introducing new ideas to [...]

Networks are often characterized by node heterogeneity for which nodes exhibit different degrees of interaction and link homophily for which nodes sharing common features tend to associate with each other. In this article, we rigorously study a directed network model that captures the former via node-specific parameterization and the latter by incorporating covariates. In particular, this model quantifies the extent of heterogeneity in terms of outgoingness and incomingness of each node by different parameters, thus allowing the number of heterogeneity parameters to be twice the number of nodes. We study the maximum likelihood estimation of the model and establish the uniform consistency and asymptotic normality of the resulting estimators. Numerical studies demonstrate our theoretical findings and two data analyses confirm the usefulness of our model. Supplementary materials for this article are available online. [ABSTRACT FROM AUTHOR]

Inappropriate methods of testing for discrimination in Supreme Court cases and in employment discrimination should be tested by formal inferential methods using p values, confidence intervals and likelihood functions.

The problem of combining graphical models of conditional independence relationships is considered in the context of cross-classified categorical data. Partial information about an underlying log-linear structure is derived from its conditional log-linear structures and used in selecting a graphical log-linear structure. The performance of the approach is illustrated in cognitive psychology applied to learning., 1. INTRODUCTION AND PROBLEM Beginning with the work of Darroch, Lauritzen, and Speed (1980), statisticians and others have begun to recognize the powerful interpretive and pedagogical value of the graphical [...]

Statistical approaches for modeling social network data have been distinguished primarily by their absence from much of the social network research literature. Researcher's work in this area is a notable exception, and thus it is a special pleasure for authors to comment on the present paper. They believe that researcher's development of exponential family models for directed graph data is a major advance that will rapidly be applied in a variety of substantive areas. In February 1978 authors first attempted to develop statistical methods for a class of log-linear models for the analysis of a social network brought by a Minnesota colleague, J. Galaskiewicz.