AN ALGORITHM FOR THE CONSTRUCTION OF OPTIMAL OR NEAR-OPTIMAL CHANGE-OVER DESIGNS

This paper presents an algorithm for the construction of optimal or near optimal change-over designs for arbitrary numbers of treatments, periods and units. Previous research on optimality has been either theoretical or has resulted in limited tabulations of small optimal designs. The algorithm consists of a number of steps:first find an optimal direct treatment effects design, ignoring residual effects, and then optimise this class of designs with respect to residual effects. Poor designs are avoided by judicious application of the (M, S)-optimality criterion, and modifications of it, to appropriate matrices. The performance of the algorithm is illustrated by examples.     

The Independence of Aptitudes,Interests, and Career Beliefs

The importance of client beliefs in career counseling depends on their ability to add unique information about the client over and above that contributed by aptitudes and interests. The Career Beliefs Inventory was administered to 200 Australian students in grade 10 together with measures of Holland's RIASEC interest themes and scholastic aptitudes. The correlations between scales from the three domains showed clearly that beliefs made a contribution distinct from that provided by aptitudes and interests. Even though the results may reflect possible sampling or method variance, career beliefs in this sample added unique information to that traditionally used in career counseling.     

A theory of policy differentiation in single issue electoral politics

Voter preferences are characterized by a parameter s (say, income) distributed on a set S according to a probability measure F. There is a single issue (say, a tax rate) whose level, b, is to be politically decided. There are two parties, each of which is a perfect agent of some constituency of voters, voters with a given value of s. An equilibrium of the electoral game is a pair of policies, b ₁ and b ₂, proposed by the two parties, such that b _i maximizes the expected utility of the voters whom party i represents, given the policy proposed by the opposition. Under reasonable assumptions, the unique electoral equilibrium consists in both parties proposing the favorite policy of the median voter. What theory can explain why, historically, we observe electoral equilibria where the ‘right’ and ‘left’ parties propose different policies? Uncertainty concerning the distribution of voters is introduced. Let {F(t)} _{t ε T} be a class of probability measures on S; all voters and parties share a common prior that the distribution of t is described by a probability measure H on T. If H has finite support, there is in general no electoral equilibrium. However, if H is continuous, then electoral equilibrium generally exists, and in equilibrium the parties propose different policies. Convergence of equilibrium to median voter politics is proved as uncertainty about the distribution of voter traits becomes small.     

Grounded Neural Networking: Modeling Complex Quantitative Data

The latest advances in artificial intelligence software (neural networking) have finally made it possible for qualitative researchers to apply the grounded theory method to the study of complex quantitative databases in a manner consistent with the postpositivistic, neopragmatic assumptions of most symbolic interactionists. The strength of neural networking for the study of quantitative data is twofold: it blurs the boundaries between qualitative and quantitative analysis, and it allows grounded theorists to embrace the complexity of quantitative data. The specific technique most useful to grounded theory is the Self‐Organizing Map (SOM). To demonstrate the utility of the SOM we (1) provide a brief review of grounded theory, focusing on how it was originally intended as a comparative method applicable to both quantitative and qualitative data; (2) examine how the SOM is compatible with the traditional techniques of grounded theory; and (3) demonstrate how the SOM assists grounded theory by applying it to an example based on our research.     

Minority Poverty in Nonmetropolitan Texas*

Abstract Despite a high prevalence of poverty among minorities in nonmetropolitan areas, research and policy concerns regarding poverty have continued focusing on metropolitan minorities. This study uses a model integrating individual, household, and structural factors to examine poverty among Latinos, blacks, and Anglos in nonmetropolitan and, for comparative purposes, metropolitan areas, using data from the 1985 Special Texas Census TDHS 1987. The findings show that minorities in nonmetropolitan areas tend to have the highest poverty rates. In addition, consistent as well as divergent patterns exist among the six ethnic-resident groups with respect to the relationships among the various individual, household, and structural factors and poverty.     

Case study in the Bayesian analysis of a cost‐effectiveness trial in the evaluation of health care technologies: Depression

We present a case study based on a depression study that will illustrate the use of Bayesian statistics in the economic evaluation of cost‐effectiveness data, demonstrate the benefits of the Bayesian approach (whilst honestly recognizing any deficiencies) with respect to frequentist methods, and provide details of using the methods, including computer code where appropriate. Copyright © 2003 John Wiley & Sons, Ltd.     

Data skeletons: simultaneous estimation of multiple quantiles for massive streaming datasets with applications to density estimation

We consider the problem of density estimation when the data is in the form of a continuous stream with no fixed length. In this setting, implementations of the usual methods of density estimation such as kernel density estimation are problematic. We propose a method of density estimation for massive datasets that is based upon taking the derivative of a smooth curve that has been fit through a set of quantile estimates. To achieve this, a low-storage, single-pass, sequential method is proposed for simultaneous estimation of multiple quantiles for massive datasets that form the basis of this method of density estimation. For comparison, we also consider a sequential kernel density estimator. The proposed methods are shown through simulation study to perform well and to have several distinct advantages over existing methods.