Topics on Dynamic Panel Data Models with Random Effects Using Semi-Parametric Bayesian Approach

A common assumption in fitting panel data models is normality of stochastic subject effects. This can be extremely restrictive, making vague most potential features of true distributions. The objective of this article is to propose a modeling strategy, from a semi-parametric Bayesian perspective, to specify a flexible distribution for the random effects in dynamic panel data models. This is addressed here by assuming the Dirichlet process mixture model to introduce Dirichlet process prior for the random-effects distribution. We address the role of initial conditions in dynamic processes, emphasizing on joint modeling of start-up and subsequent responses. We adopt Gibbs sampling techniques to approximate posterior estimates. These important topics are illustrated by a simulation study and also by testing hypothetical models in two empirical contexts drawn from economic studies. We use modified versions of information criteria to compare the fitted models.     

Marginalized random-effects models for clustered binomial data through innovative link functions

AStA Advances in Statistical Analysis - Random-effects models are frequently used to analyze clustered binomial data. The direct computation of the marginal mean response, when integrated over the...     

The Public Sector and income distribution in South Africa during the period 1970–82

This study considers the impact of fiscal and monetary policy on the distribution of income in South Africa. It analyses expenditure using two approaches: (1) the general approach which considers the extent of programmes having a relatively high distributive contact in favour of the poor and (2) the direct money‐flow approach which assumes that expenditure benefits those who receive it as their income. Although neither approach is perfect, it can be concluded that the general composition of expenditure has favoured the poor. It is also argued that the tax system is partly progressive and partly regressive and that it has become more regressive, although remaining in balance progressive. Monetary policy during the 1970s, however, has had definite regressive income distribution effects.     

Bayesian analysis of longitudinal ordered data with flexible random effects using McMC: application to diabetic macular Edema data

In the analysis of correlated ordered data, mixed-effect models are frequently used to control the subject heterogeneity effects. A common assumption in fitting these models is the normality of random effects. In many cases, this is unrealistic, making the estimation results unreliable. This paper considers several flexible models for random effects and investigates their properties in the model fitting. We adopt a proportional odds logistic regression model and incorporate the skewed version of the normal, Student's t and slash distributions for the effects. Stochastic representations for various flexible distributions are proposed afterwards based on the mixing strategy approach. This reduces the computational burden being performed by the McMC technique. Furthermore, this paper addresses the identifiability restrictions and suggests a procedure to handle this issue. We analyze a real data set taken from an ophthalmic clinical trial. Model selection is performed by suitable Bayesian model selection criteria.     

A comparative study on estimation methods to deal with the endogeneity in linear random-intercept models with an extension

In this paper, we investigate estimation methods to deal with situations where random intercepts are associated to time-varying covariates in the context of linear mixed models. First, a review of previous ways to deal with this so-called endogeneity issue is presented, then a new method based on shared random effects is proposed. Simulation studies and an empirical example are utilized to assess the performance of our proposed method. It is shown that our new approach is more efficient than most competitors and is robust to the misspecification of the random-effects distributions.     

A Multilayered Box Model for Calculating Preliminary Remediation Goals in Soil Screening

In the process of screening a soil against a certain contaminant, we define the health-risk-based preliminary remediation goal (PRG) as the contaminant concentration above which some remedial action may be required. Thus, PRG is the first standard (or guidance) for judging a site. An overestimated PRG (a too-large value) may cause us to miss some contaminated sites that can threaten human health and the environment. An underestimated PRG (a too-small value), on the other hand, may lead to unnecessary cleanup and waste tremendous resources. The PRGs for soils are often calculated on the assumption that the contaminant concentration in soil does not change with time. However, that concentration usually decreases with time as a result of different chemical and transport mechanisms. The static assumption thus exaggerates the long-term exposure dose and results in a too-small PRG. We present a box model that considers all important transport processes and obeys the law of mass conservation. We can use the model as a tool to estimate the transient contaminant concentrations in air, soil, and ground water. Using these concentrations in conjunction with appropriate health-risk parameters, we may estimate the PRGs for different contaminants. As an example, we calculated the tritium PRG for residential soils. The result is quite different from, but within the range of, the two versions of the corresponding PRG previously recommended by the U.S. EPA.