Doubly adaptive biased coin designs with heterogeneous responses

Doubly adaptive biased coin design (DBCD) is an important family of response-adaptive randomization procedures for clinical trials. It uses sequentially updated estimation to skew the allocation probability to favor the treatment that has performed better thus far. An important assumption for the DBCD is the homogeneity assumption for the patient responses. However, this assumption may be violated in many sequential experiments. Here we prove the robustness of the DBCD against certain time trends in patient responses. Strong consistency and asymptotic normality of the design are obtained under some widely satisfied conditions. Also, we propose a general weighted likelihood method to reduce the bias caused by the heterogeneity in the inference after a trial. Some numerical studies are also presented to illustrate the finite sample properties of DBCD.     

Declining inter-industry wage dispersion in the US

Industrial effects have long been significant factors in wage inequality. Previous research indicates that wage differentials across industries were increasing through the mid 1980s. Using more recent data, however, we find that the level of inter-industry wage dispersion declined by 36% from 1986 to 2002 despite the continued trend towards increasing inequality in the labor force. This decline in inter-industry wage dispersion is evident across gender and educational groups. Using multilevel growth curve models, our multivariate results indicate that the decline is only weakly related to industrial changes in education, occupation or even productivity despite the fact that the latter variable had been a critical factor in the prior period. Indicators of globalization and downsizing also do not appear to explain this decline. For the more recent period, the most important factors associated with the narrowing of inter-industry wage dispersion are reduced unionization rates and the higher proportion of casual workers. We interpret these results as suggesting that firms may now be less economically obliged to pass on a portion of their rents to broad groups of workers and may instead be engaged in more idiosyncratic processes of negotiation with individual workers based on micro-level sources of bargaining power.     

Variance estimation for the 2006 French housing survey

In the 2006 French housing survey, information is collected on many aspects of housing to describe the housing stock in France and the housing conditions of French households. The basic national sample results from a multistage sampling design. Complementary samples were selected to perform accurate estimations for socio-demographic domains. Some French regions proceeded to a regional and local extension of the national sample. The variance is estimated for a region with a regional and local extension of the basic national sample.     

Synthetic double sampling s chart

This study proposes a synthetic double sampling s chart that integrates the double sampling (DS) s chart and the conforming run length chart. An optimization procedure is proposed to compute the optimal parameters of the synthetic DS s chart. The performance of the synthetic DS s chart is compared with other existing control charts for monitoring process standard deviation. The results show that the synthetic DS s chart is more effective for detecting increases in the process standard deviation for a wide range of shifts. An example is provided to illustrate the operation procedure of the synthetic DS s chart.     

Estimation of finite population mean in simple and stratified random sampling using two auxiliary variables

We propose an improved difference-cum-exponential ratio type estimator for estimating the finite population mean in simple and stratified random sampling using two auxiliary variables. We obtain properties of the estimators up to first order of approximation. The proposed class of estimators is found to be more efficient than the usual sample mean estimator, ratio estimator, exponential ratio type estimator, usual two difference type estimators, Rao (1991) estimator, Gupta and Shabbir (2008) estimator, and Grover and Kaur (2011) estimator. We use six real data sets in simple random sampling and two in stratified sampling for numerical comparisons.     

Use of online crowdsourcing platforms for gambling research

Crowdsourcing platforms like Amazon’s Mechnical Turk and Crowdflower have been touted to be a cost-effective way to collect large amounts of behavioural data. Across four large-n studies, gambling-related behaviours, tendencies and traits among participants in these labour markets were examined. In Studies 1 and 2, both conducted on Crowdflower, problem gamblers (as measured by the benchmark Problem Gambling Severity Index) comprised 24.5% and 21.9% of participants, respectively. In Study 3, conducted on Mechanical Turk, problem gamblers comprised 9.0% of participants. In Study 4, a two-wave longitudinal study conducted on Crowdflower, problem gamblers comprised 13.5% of participants in wave one and 14.8% of participants in wave two. In Studies 2 and 3, strong convergent associations were demonstrated across various measures of problem gambling tendencies and general gambling involvement. Furthermore, it was demonstrated that gambling was associated with personality traits (impulsivity, sensation-seeking, self-control), risk attitudes, affect, and behavioural risk-taking consistent with previous research. In Study 4, it was demonstrated that measures of problem gambling have acceptable test-retest reliability. Online crowdsourcing platforms appear to offer access to samples with remarkably high proportions of problem gamblers. However, this characteristic means that such samples are not necessarily representative of gambling tendencies among more general populations.     

Adaptive multiple importance sampling for Gaussian processes

In applications of Gaussian processes (GPs) where quantification of uncertainty is a strict requirement, it is necessary to accurately characterize the posterior distribution over Gaussian process covariance parameters. This is normally done by means of standard Markov chain Monte Carlo (MCMC) algorithms, which require repeated expensive calculations involving the marginal likelihood. Motivated by the desire to avoid the inefficiencies of MCMC algorithms rejecting a considerable amount of expensive proposals, this paper develops an alternative inference framework based on adaptive multiple importance sampling (AMIS). In particular, this paper studies the application of AMIS for GPs in the case of a Gaussian likelihood, and proposes a novel pseudo-marginal-based AMIS algorithm for non-Gaussian likelihoods, where the marginal likelihood is unbiasedly estimated. The results suggest that the proposed framework outperforms MCMC-based inference of covariance parameters in a wide range of scenarios.     

Model comparison of linear and nonlinear Bayesian structural equation models with dichotomous data

In this article, dichotomous variables are used to compare between linear and nonlinear Bayesian structural equation models. Gibbs sampling method is applied for estimation and model comparison. Statistical inferences that involve estimation of parameters and their standard deviations and residuals analysis for testing the selected model are discussed. Hidden continuous normal distribution (censored normal distribution) is used to solve the problem of dichotomous variables. The proposed procedure is illustrated by a simulation data obtained from R program. Analyses are done by using R2WinBUGS package in R-program.     

Monitoring the coefficient of variation using a variable sample size and sampling interval control chart

This article proposes a CV chart by using the variable sample size and sampling interval (VSSI) feature to improve the performance of the basic CV chart, for detecting small and moderate shifts in the CV. The proposed VSSI CV chart is designed by allowing the sample size and the sampling interval to vary. The VSSI CV chart's statistical performance is measured by using the average time to signal (ATS) and expected average time to signal (EATS) criteria and is compared with that of existing CV charts. The Markov chain approach is employed in the design of the chart.