Latent classes and group membership

A technique for analyzing group membership data, such as interlocking directorates, based on the assumption of latent classes of individuals, is described and illustrated with two data sets. The technique partitions individuals and/or groups into homogeneous sets and can be used to create measures of structural centrality for groups and for individuals.     

Statistical and Historical Analyses of Nations Which Deviate From the Size-Density Law

We examine nations which deviate significantly from the ?2/3 sizedensity slope previously observed for the aggregated political subdivisions of 98 modern nations and previously derived from the theory of timeminimization. We correct Stephan’s original list of deviant nations, demonstrate that all slopes more negative than ?2/3 were due to the erroneous inclusion of cities in the original data sets, develop a theory to account for the less negative deviant slopes, and test the theory both through statistical manipulation of current data and through analysis of available historical data.     

An infinite class of PBIB designs

In this paper, an infinite class of partially balanced incomplete block (PBIB) designs of m+1 associate classes is constructed through the use of a series of row-orthogonal matrices known as partially balanced orthogonal designs (PBOD) of m-associate classes. For the purpose, a series of PBOD is obtained through a method described herein. An infinite class of regular GD designs is also reported.     

The end of rapid increase in the use of oral anovulants? Some problems in the interpretation of time series of oral use among married women

Several problems in the interpretation of time series of oral anovulant usage are discussed, among them the shifting marriage-duration distribution over time. Standardization fails to negate the conclusion that the rate of increase in the proportion of women using orals has slackened since January 1966. Separate analyses by duration-specific rates of use and by marriage cohorts show that the two factors sustaining the initial spectacular growth rate in the use of orals-the ever—higher initial use rates of new marriage cohorts and the rapid adoption of orals among earlier cohorts at later durations—have lost their sustaining force. Further analysis indicates that concern over the pill as a health hazard is a major deterrent to substantial increase in oral usage. A final distribution of women by risk status with respect to oral use, and current and prospective oral use, shows that under present circumstances the maximum oral use rate will be considerably less than unity.