Residual analysis for parametric models fit to censored data

The focus of this paper is on residual analysis for the lognormal and extreme value or Weibull models, although the proposed methods can be applied to any parametric model. Residuals developed by Barlow and Prentice (1988) for the Cox proportional hazards model are extended to the parametric model setting. Three different residuals are proposed based on this approach with two residuals measuring the impact of survival time and one measuring the impact of the covariates included in the model. In addition, a residual derived from the deviations equality presented in Efron and Johnstone (1990) and the residual proposed by Joergensen (1984) for censored data models are discussed.     

Indicators of Adolescent Depression and Relationship Progression in Emerging Adulthood

Adolescent depression may be associated with future relationship problems that have long‐term consequences given the developmental importance and health benefits of forming committed unions in emerging adulthood. The authors examined associations between emotional and behavioral indicators of adolescent depression (depressive symptoms, alcohol problems, and suicidal ideation) and romantic relationship and union formation and dissolution in emerging adulthood (n = 14,146) using the National Longitudinal Study of Adolescent Health. Adolescent alcohol problems were associated with more romantic relationships in emerging adulthood. Emerging adults with depressive symptoms or alcohol problems in adolescence were significantly more likely to enter into a cohabiting union, and those with adolescent alcohol problems were less likely to marry. Cohabiting emerging adults with a history of adolescent depressive symptoms were less likely to marry, whereas suicidal ideation was associated with a decreased likelihood of cohabitation dissolution. Implications for future research are discussed.     

Complex archetypal structures that underlie the ‘human condition’

Abstract

In this article I synthesise notions from several different fields to argue that there are archetypal structures that underlie the human condition that can best be understood through the language of complexity. In so doing I accommodate the disparity that exists between different branches of the social sciences through the provision of a holistic overview that accords with the idea of a coherent scientific body of knowledge. To do this I use a Jungian typology to reinforce the parameters of four archetypes by which organisation and individuals have been interpreted, and locate this within some findings from evolutionary psychology, thereby building a meta-typology of the human condition. I conclude by suggesting that this can best be understood through the notions of relationship and knowledge through the opposite, and by examining some of the implications this approach holds for the field.     

On confident men and rational women: It’s all on your mind(set)

We tested the hypothesis that inducing the deliberative and the implemental mindset differently affects judgment and decision making. More specifically, we explored mindset effects on decision makers’ confidence ratings, risk preferences, and susceptibility to anchoring effects. As earlier research on mindsets showed that individual differences sometimes moderate mindset effects, we also tested for interaction effects of mindset and gender. For confidence ratings, we found a main effect of mindset and a main effect of gender. For risk preferences and anchoring effects, mindset interacted with gender. When being in an implemental mindset, the judgments of female decision makers came closer to their actual performance compared to being in a deliberative mindset where they were observed as underconfident. Male decision makers were already overconfident in the deliberative mindset and showed even more overconfidence when being in an implemental mindset. Concerning risk attitudes it was found that female decision makers were more prone to choose the less risky, but also less profitable option (in terms of expected payoffs) when they were in the deliberative compared to the implemental mindset. For men the opposite effects were observed. When investigating anchoring effects, male but not female participants’ judgments were influenced by mindset: In an implemental mindset, male participants followed an irrelevant anchor more strongly (i.e., made more anchor-consistent judgments) compared to being in a deliberative mindset.     

Continuous processes derived from the solution of generalized Langevin equation: theoretical properties and estimation

ABSTRACT

In this paper we present a class of continuous-time processes arising from the solution of the generalized Langevin equation and show some of its properties. We define the theoretical and empirical codifference as a measure of dependence for stochastic processes. As an alternative dependence measure we also consider the spectral covariance. These dependence measures replace the autocovariance function when it is not well defined. Results for the theoretical codifference and theoretical spectral covariance functions for the mentioned process are presented. The maximum likelihood estimation procedure is proposed to estimate the parameters of the process arising from the classical Langevin equation, i.e. the Ornstein–Uhlenbeck process, and of the so-called Cosine process. We also present a simulation study for particular processes arising from this class showing the generation, and the theoretical and empirical counterpart for both codifference and spectral covariance measures.     

Clustering time-course microarray data using functional Bayesian infinite mixture model

This paper presents a new Bayesian, infinite mixture model based, clustering approach, specifically designed for time-course microarray data. The problem is to group together genes which have “similar” expression profiles, given the set of noisy measurements of their expression levels over a specific time interval. In order to capture temporal variations of each curve, a non-parametric regression approach is used. Each expression profile is expanded over a set of basis functions and the sets of coefficients of each curve are subsequently modeled through a Bayesian infinite mixture of Gaussian distributions. Therefore, the task of finding clusters of genes with similar expression profiles is then reduced to the problem of grouping together genes whose coefficients are sampled from the same distribution in the mixture. Dirichlet processes prior is naturally employed in such kinds of models, since it allows one to deal automatically with the uncertainty about the number of clusters. The posterior inference is carried out by a split and merge MCMC sampling scheme which integrates out parameters of the component distributions and updates only the latent vector of the cluster membership. The final configuration is obtained via the maximum a posteriori estimator. The performance of the method is studied using synthetic and real microarray data and is compared with the performances of competitive techniques.     

Nonparametric Tests for Comparing Several Coefficients of Variation

We consider the problem of comparing (k + 1) coefficients of variation. We are interested in testing the null hypothesis that the coefficients of variation are equal against each of the alternatives: (a) some populations have different coefficients of variation and (b) the coefficients of variation are ordered. Three nonparametric test statistics are proposed and their asymptotic theory is developed. We compared the proposed tests together with another parametric test using two Monte Carlo studies to estimate their probabilities of Type I error and powers. An illustration of the proposed tests using a real data set is given.     

Identification of multiple influential observations in logistic regression

The identification of influential observations in logistic regression has drawn a great deal of attention in recent years. Most of the available techniques like Cook's distance and difference of fits (DFFITS) are based on single-case deletion. But there is evidence that these techniques suffer from masking and swamping problems and consequently fail to detect multiple influential observations. In this paper, we have developed a new measure for the identification of multiple influential observations in logistic regression based on a generalized version of DFFITS. The advantage of the proposed method is then investigated through several well-referred data sets and a simulation study.     

The Steady-State Performance of Cumulative Count of a Conforming Control Chart Based On Runs Rules

Cumulative count of conforming control chart is usually used to monitor fraction nonconforming in high-yield processes. In this article, we propose m-of-m control chart based on cumulative count of conforming units for high-yield processes. The steady-state properties of the m-of-m control chart are investigated. We compare performance of the m-of-m control chart with control chart based on cumulative count of conforming units. We present Markov chain model of the m-of-m control chart to evaluate average run length, standard deviation of run length and quartiles.     

Empirical Likelihood Inference for the Cox Model with Time-dependent Coefficients via Local Partial Likelihood

Abstract.  The Cox model with time-dependent coefficients has been studied by a number of authors recently. In this paper, we develop empirical likelihood (EL) pointwise confidence regions for the time-dependent regression coefficients via local partial likelihood smoothing. The EL simultaneous confidence bands for a linear combination of the coefficients are also derived based on the strong approximation methods. The EL ratio is formulated through the local partial log-likelihood for the regression coefficient functions. Our numerical studies indicate that the EL pointwise/simultaneous confidence regions/bands have satisfactory finite sample performances. Compared with the confidence regions derived directly based on the asymptotic normal distribution of the local constant estimator, the EL confidence regions are overall tighter and can better capture the curvature of the underlying regression coefficient functions. Two data sets, the gastric cancer data and the Mayo Clinic primary biliary cirrhosis data, are analysed using the proposed method.