Some simple log-linear models for the analysis of ordinal response data

A family of log-linear models are proposed to describe contingency tables in which one variable can be considered as the response to the remaining. The proposed models take into account the ordering nature of the response categories and have structure similar to that employed in polynomial regression. Stochastic ordering of the response distributions under the proposed models is discussed and the model-reduction techniques are developed. The proposed models are applied to two data sets previously analysed in the literature.     

The decision to add a second system producer

Typically in the United States only one contractor (the developer or first producer) makes a weapon system because it is too expensive or for some other reason infeasible to develop a second source. Congress and higher headquarters are constantly challenging the Armed Services to add other sources who can competitively make their systems. This paper describes a method of analyzing this complex problem which often has very large monetary and other significant impacts. Procedures and a case study are presented which consider cost and other criteria which might be used to make the second source decision.     

THE DEMOGRAPHICS OF A NEW VOTING GAP MARITAL DIFFERENCES IN AMERICAN VOTING

There is now a marriage gap in presidential voting in the UnitedStates, with married people voting about 10%–15% moreRepublican than unmarried people. This voting difference developedin the 1972 election, and it is also evident in party identification.The proportion of unmarried people has doubled since 1964, sothis cleavage is important. The gap can be explained statisticallyin terms of race and income, but still may give politiciansthe opportunity to gain votes by appealing to voters on thebasis of family issues.     

A Quantitative Estimate of Leukemia Mortality Associated with Occupational Exposure to Benzene3

In 1980, the U.S. Supreme Court vacated a revised occupational standard for benzene, stating that the Occupational Safety and Health Administration (OSHA) had failed to demonstrate that significant health risks existed under the current standard. This decision has been interpreted by OSHA as requiring the consideration of quantitative risk assessments, whenever possible, in the development of regulations for occupational carcinogens. In light of this decision, the available epidemiologic evidence was used to generate a quantitative risk assessment for benzene. Uncertainties regarding the levels and lengths of benzene exposure for the studied cohorts were incorporated into the analysis. Based on the one-hit model, the assessment indicates that a working lifetime exposure to benzene at the current permissible exposure level (10 ppm) poses a substantial excess risk of death from leukemia. This report discusses the calculation of the risk estimates, the basis for relying on certain assumptions, and the inherent limitations of using epidemiologic studies to quantify cancer risks.     

Responsibility for Feelings: A Multilateral Approach to Causation

The multilateral approach to causation posits four possible causes which need to be considered in the assessment and treatment of discord between partners in a primary relationship. These causes include the precipitating event, the reaction of a person to the precipitating event, the interactional dynamics of the partners, and the goals and objectives of the relationship which are being frustrated. A multilateral approach permits a person to accept his or her own legitimate responsibility for behavior and feelings without a responsibility "underload" or "overload." The use of multilateral causation is intended as a therapeutic safeguard against oversimplification.     

A mixture model for random graphs

The Erdös–Rényi model of a network is simple and possesses many explicit expressions for average and asymptotic properties, but it does not fit well to real-world networks. The vertices of those networks are often structured in unknown classes (functionally related proteins or social communities) with different connectivity properties. The stochastic block structures model was proposed for this purpose in the context of social sciences, using a Bayesian approach. We consider the same model in a frequentest statistical framework. We give the degree distribution and the clustering coefficient associated with this model, a variational method to estimate its parameters and a model selection criterion to select the number of classes. This estimation procedure allows us to deal with large networks containing thousands of vertices. The method is used to uncover the modular structure of a network of enzymatic reactions.