John Deegan

On the Occurrence of Standardized Regression Coefficients Greater than One.

For some investigators, the occurrence of standardized regression coefficients greater than one in a model raises questions concerning the legitimacy of such coefficients, and poses serious problems of interpretation (particularly for those employing path analytic procedures). It is demonstrated here that standardized regression coefficients greater than one can legitimately occur. Furthermore, the relationship between the occurrence of such coefficients and the extent of multicollinearity present among the set of predictor variables in an equation is examined. Comments on the interpretation of these coefficients, including a discussion of their interpretation in the context of path analysis, are provided.

A Theory of Political Districting

Fair districting requires more than compact, contiguous equal-sized districts; namely, sets of districts should also possess certain features. Specifically, they should be neutral (treat all parties alike) and responsive to changes in votes. In order to establish the extent to which these goals can be achieved, we give a precise definition to the concept of neutrality and expand the notion of responsiveness into three characteristics: the range in actual votes over which a districting plan is responsive; the degree of responsiveness in the vicinity of the “normal vote” (i.e., competitiveness); and the constancy of the swing ratio (i.e., the rate at which vote changes yield seat changes) over a range of votes. We show that while all possible values for these features are readily attainable when considered individually, certain combinations of values cannot be achieved. Finally, we identify the nature of the compromises required and the properties that the compromises possess, and show the kinds of trade-offs that result in reasonably fair districting plans.

An Axiomated Family of Power Indices for Simple n -Person Games

In this probabilistic generalization of the Deegan-Packel power index, a new family of power indices based on the notions of minimal winning coalitions and equal division of pay offs is developed. The family of indices is parameterized by allowing minimal winning coalitions to form in accordance with varying probability functions. These indices are axiomatically characterized and compared to other similarly characterized indices. Additionally, a dual family of minimal blocking coalition indices and their characterization axioms is presented.

The Process of Political Development

Although a number of investigators have attempted to identify empirically a process of political development, substantial controversy still surrounds a determination of the causal factors involved. It is my contention that this state of affairs is the result of inadequacies inherent in traditional techniques of causal modeling, and aggravated when multicollinear variables are involved. To resolve this problem I first review a technique capable of reducing the confounding effects of multicollinearity. I then illustrate use of this technique, as well as a strategy for inferring causal relationships, by means of a reanalysis of published data used to construct models of political development. The strategy for causal inference utilized herein is derived from knowledge of the effects of model specification errors. On the basis of these findings a new causal model of political development, which is both theoretically and empirically consistent, is presented.

Specification error in causal models

A typology of model specification errors, initially devised for single-equation multiple-regression models, is extended to include systems of equations embodied in recursive causal models. Analytical results for the error forms identified are presented. On the basis of these results a number of inadequacies of traditional techniques of causal inference are identified. In place of the traditional techniques of causal inference an alternative strategy for constructing recursive causal models, based on specification error knowledge, is offered. An illustrative application is also included.

Looking Back At Love Canal

The comprehensive environmental study which describes the results of a monitoring program conducted by EPA at Love Canal is evaluated by EPAs former study director. Attention is focused on the episodes history and the agencys study methods. The aim of the program was to constitute a study team, design a monitoring study, reprogram and reallocate the financial resources needed to conduct the study, and identify and employ contractors who would collect and analyze environmental samples. The agency was directed to ensure the quality of the data acquired from various environmental media and analyzed by numerous laboratories; to integrate, interpret, and report the data; and to assess, from an environmental perspective, the habitability of the area.

The Consequences of Model Misspecification in Regression Analysis.

In ordinary least squares regression analysis the desired property of unbiasedness in estimated coefficients is contingent upon the correspondence of the fitted model with the true underlying data generating process. This paper focuses on developing a systematic characterization of the error forms resulting from model misspecification in single equation models. The consequences of model misspecification, for the error forms identified, are also evaluated.

Layered chalcogenide glass structures for IR lenses

A technique for fabricating novel infrared (IR) lenses can enable a reduction in the size and weight of IR imaging optics through the use of layered glass structures. These structures can range from having a few thick glass layers, mimicking cemented doublets and triplets, to having many thin glass layers approximating graded index (GRIN) lenses. The effectiveness of these structures relies on having materials with diversity in refractive index (large Δn) and dispersion and similar thermo-viscous behavior (common glass transition temperature, ΔTg = 10°C). A library of 13 chalcogenide glasses with broad IR transmission (NIR through LWIR bands) was developed to satisfy these criteria. The lens fabrication methodology, including glass design and synthesis, sheet fabrication, preform making, lens molding and surface finishing are presented.

An analysis of nonpartisan election media expenditure decisions using limited dependent variable methods

Abstract In order to maximize the information disseminating function of campaigning, candidates for political office must make numerous decisions regarding the allocation of resources for various campaign activities. One of the most important financial decisions a candidate must make occurs in determining the extent and character of mass media expenditures. To determine mass media budget allocations, two decisions are required. First, the type(s) of mass media to employ in the campaign, and second, the amount of the budget to allocate to each media selected. These two decisions are not necessarily independent since high fixed costs are associated with employing electronic media. In this paper, we examine one particular form of mass media, television, and its use in nonpartisan elections. Because of the nature of the decision process in using television advertising an estimation technique for limited dependent variables is employed to predict first, the probability that a candidate will use television advertising in a campaign, and second, the amount spent on such advertising. The occurrence of these outcomes is considered to be a function of both the total financial resources available to the candidate and the type (importance) of the political office sought.

A test of the numerical accuracy of some matrix inversion algorithms commonly used in least squares programs

In recent years a number of investigations of the accuracy of least squares programs found gross numerical inaccuracies in many programs. Apparently those programs employing elimination algorithms for matrix inversion fared the poorest. Our findings do not support this conclusion. Our findings demonstrate that elimination algorithms, when used for solving least squares problems, are not intrinsically unstable. A brief discussion of an appropriate strategy to follow for estimating least squares coefficients is included.