Communications in Statistics - Theory and Methods最新文献

The multinomial probit model has been a prominent tool to analyze nominal categorical data, but the computational complexity of maximum likelihood functions presents challenges in the usage of this model. Furthermore, the model identification is extremely tenuous and usually necessitates the covariance matrix of the latent multivariate normal variables to be a restricted covariance matrix, which brings a rigorous task for both likelihood-based estimation and Markov chain Monte Carlo (MCMC) sampling. We tackle this issue by constructing a non-identifiable model and developing parameter-expanded data augmentation. Our proposed methods circumvent sampling a restricted covariance matrix commonly implemented by a painstaking Metropolis-Hastings (MH) algorithm and enable to sample a covariance matrix without restriction through a Gibbs sampler. Therefore, our proposed methods advance the convergence and mixing of the MCMC components considerably. We investigate our proposed methods along with the method based on the identifiable model through simulation studies and further illustrate their performance by an application to consumer choice on liquid laundry detergents data.

Statistical methods have been developed for regression modeling of the cumulative incidence function (CIF) given left-truncated right-censored competing risks data. Nevertheless, existing methods typically involve complicated weighted estimating equations or nonparametric conditional likelihood function and often require a restrictive assumption that censoring and/or truncation times are independent of failure time. The pseudo-observation (PO) approach has been used in regression modeling of CIF for right-censored competing risks data under covariate-independent censoring or covariate-dependent censoring. We extend this approach to left-truncated right-censored competing risks data and propose to directly model the CIF based on POs, under general truncation and censoring mechanisms. We adjust for covariate-dependent truncation and/or covariate-dependent censoring by incorporating covariate-adjusted weights into the inverse probability weighted (IPW) estimator of the CIF. We derive large sample properties of the proposed estimators under reasonable model assumptions and regularity conditions and assess their finite sample performances by simulation studies under various scenarios. We apply the proposed method to a cohort study on pregnancy exposed to coumarin derivatives.

Count outcomes often occur in cluster randomized trials. Particularly in the context of epidemiology, the ratio of incidence rates has been used to assess the effectiveness of an intervention. In practice, cluster sizes typically vary across clusters, and sample size estimation based on a constant cluster size assumption may lead to underpowered studies. To address this issue, we propose a sample size method based on the generalized estimating equation (GEE) approach to test the ratio of two incidence rates. A closed-form sample size formula is presented, which is flexible to account for unbalanced randomization and randomly varying cluster sizes. Simulations were performed to assess its performance. In cluster randomized trials of vaccine efficacy, the ratio of disease incidence rates has been frequently used to demonstrate that the vaccine reduces the occurrence of a disease compared to placebo or active control. An application example to the design of a vaccine efficacy cluster randomized trial is presented.

Multivariate probit models have been popularly utilized to analysis multivariate ordinal data. However, the identifiable multivariate probit models entail the covariance matrix for the underlying multivariate normal variables to be a correlation matrix, which brings a rigorous task to conduct efficient statistical analysis. Parameter expansion to make the identifiable model to be non-identifiable has been inevitably explored. However, the effect of the expanded parameters on the convergence of Markov chain Monte Carlo (MCMC) is seldomly investigated; in addition, the comparison of MCMC developed based on the identifiable model and that based on the non-identifiable model is hardly ever explored, especially for data with large sample sizes. In this paper, we conduct a thorough investigation to illustrate the effect of the expanded parameters on the convergence of MCMC and compare the behavior of MCMC between the identifiable and non-identifiable models. Our investigation provides a practical guide regarding the construction of non-identifiable models and development of corresponding MCMC sampling methods. We conduct our investigation using simulation studies and present an application using data from the Russia Longitudinal Monitoring Survey-Higher School of Economics (RLMS-HSE) study.