Analysis of a marker for cancer of the thyroid with a limit of detection

Limit of detection (LoD) is a common problem in the analysis of data generated by instruments that cannot detect very small concentrations or other quantities, resulting in left-censored measurements. Methods intended for data that are not subject to this problem are often difficult to modify for censoring. We adapt the simulation-extrapolation method, devised originally for fitting models with measurement error, to dealing with LoD in conjunction with a mixture analysis. The application relates the levels of thyroglobulin in individuals with cancer of the thyroid before and after treatment with radioactive iodine I–131. We conclude that the fitted mixture components correspond to levels of effectiveness of the treatment.     

Correlated defaults,temporal correlation,expert information and predictability of default rates

ABSTRACT

Dependence among defaults both across assets and over time is an important characteristic of financial risk. A Bayesian approach to default rate estimation is proposed and illustrated using prior distributions assessed from an experienced industry expert. Two extensions of the binomial model are proposed. The first allows correlated defaults yet remains consistent with Basel II’s asymptotic single-factor model. The second adds temporal correlation in default rates through autocorrelation in the systemic factor. Implications for the predictability of default rates are considered. The single-factor model generates more forecast uncertainty than does the parameter uncertainty. A robustness exercise illustrates that the correlation indicated by the data is much smaller than that specified in the Basel II regulations.     

Waiting Line Effects in Analytical Maintenance Float Policy

The equipment failure distributions commonly identified in practice pose great difficulties in the establishment of sound maintenance total float policy. Some of the existing analytical constructs utilize the reliability ratio of the equipment in operations in order to bypass the obstacles and arrive at applicable solutions. In this paper we give a summary of these theoretical models. In addition, we provide insight by testing the significance of factors utilized in maintenance total float determination. The paper examines the effect of the assumption made by the analytical model of no waiting time for repair. Our test showed that there is no statistically significant difference between the analytical model and simulation when waiting is allowed for the normal and uniform distributions. The analytical model thus offers some advantage in its use to estimate maintenance float due to its simplicity. When failure distribution is an exponential, Erlang-2 or lognormal, the assumption of no waiting time for repair must be kept in order to use the analytical model.     

Mobilized by Direct Democracy: Short‐Term Versus Long‐Term Effects and the Geography of Turnout in Ballot Measure Elections*

Objectives. A number of recent studies find that direct democracy increases voter turnout. In this article, we ask: Who does direct democracy mobilize to vote and how are they mobilized? We distinguish between long‐term and short‐term effects on voter turnout, noting that much of the current literature has focused on participatory theory. Methods. Our research design harnesses the power of geographic information systems and examines turnout in special initiative‐only elections using registered voter lists. Our model draws on individual and Census tract data, incorporated using a hierarchical generalized linear model. Results. The findings demonstrate how partisan context mitigates the potential for direct democracy to mobilize from the middle, and clarifies the dominance of short‐term as opposed to long‐term effects in increasing voter participation in ballot initiative elections. Conclusion. Mobilization via direct legislation occurs mostly because voters are actively mobilized by partisan campaigns, not because of an increase in participatory fervor.     

Singly and Doubly Censored Current Status Data: Estimation, Asymptotics and Regression

In biostatistical applications interest often focuses on the estimation of the distribution of time between two consecutive events. If the initial event time is observed and the subsequent event time is only known to be larger or smaller than an observed point in time, then the data is described by the well-understood singly censored current status model, also known as interval censored data, case I. Jewell et al. (1994) extended this current status model by allowing the initial time to be unobserved, with its distribution over an observed interval [A, B] known; the data is referred to as doubly censored current status data. This model has applications in AIDS partner studies. If the initial time is known to be uniformly distribute d, the model reduces to a submodel of the current status model with the same asymptotic information bounds as in the current status model, but the distribution of interest is essentially the derivative of the distribution of interest in the current status model. As a consequence the non-parametric maximum likelihood estimator is inconsistent. Moreover, this submodel contains only smooth heavy tailed distributions for which no moments exist. In this paper, we discuss the connection between the singly censored current status model and the doubly censored current status model (for the uniform initial time) in detail and explain the difficulties in estimation which arise in the doubly censored case. We propose a regularized MLE corresponding with the current status model. We prove rate results, efficiency of smooth functionals of the regularized MLE, and present a generally applicable efficient method for estimation of regression parameters, which does not rely on the existence of moments. We also discuss extending these ideas to a non-uniform distribution for the initial time.     

The Impact of In-service Training within Social Services

Summary In-service training within social service agencies is recognizedas a key means through which staff are provided with the necessaryknowledge and skills to improve overall agency performance andachieve the objectives of social policy. Furthermore, reportsof training expenditure within social services departments inthe UK suggest increasing amounts of monies are invested insuch activity in order to meet the changing demands placed onsocial care and greater expectations for higher standards inservice delivery. Yet to what extent is the faith placed inmuch of this training by social service agencies actually justified?In short, how much do we know about the actual impact of in-servicetraining within social service organizations in terms of itachieving the aims set by policy makers? This article attemptsto answer this question in order to assist in the developmentof a far more empirically based framework for understandingtraining processes within social service agencies. A reviewof the literature for studies published between 1974 and 1997detailing an evaluation of in-service training programmes withinsocial service agencies identified only 20 such studies. Problemsassociated with both the evaluative criteria utilized and theresearch methodologies employed in these studies meant thatin many instances conclusions regarding the actual impact oftraining could only be tentatively judged. As a result, an analysisof these studies found that although training may have an impacton trainees in terms of satisfaction or knowledge gain, resultsregarding impact on behaviour are far more inconclusive. Inaddition it is not at all certain that such training will necessarilyresult in changes in performance back in the workplace. Thefindings from the review suggest far more research is requiredof sufficient rigour to underpin our knowledge in this importantarea.