Parameter estimation of Poisson generalized linear mixed models based on three different statistical principles : a simulation study

Generalized linear mixed models are flexible tools for modeling non-normal data and are useful for accommodating overdispersion in Poisson regression models with random effects. Their main difficulty resides in the parameter estimation because there is no analytic solution for the maximization of the marginal likelihood. Many methods have been proposed for this purpose and many of them are implemented in software packages. The purpose of this study is to compare the performance of three different statistical principles - marginal likelihood, extended likelihood, Bayesian analysis-via simulation studies. Real data on contact wrestling are used for illustration.

Att integrera i det svenska samhället : En undersökning om kvinnor med invandrarbakgrund och deras syn på introduktion och andra faktorer som påverkar inträdet i det svenska arbetslivet och samhället

The integration policy in Sweden shall encourage individuals to support themselves and take part in society. It shall alsocontribute to equal rights and opportunities for women as well as for men. In Borlänge this has resulted in a program ofintroduction for new arrivals from other countries. However, at the unit responsible for economic support, they havediscovered that women with immigrant background more often than men seem to have trouble starting or became to anending of the program, which then especially leads women to a long-term dependence for economic support. The purpose of this study has therefore been to investigate what factors affect immigrant women’s participation in theintroduction, and what significance this participation has for their possibilities to become economically self-supporting,and integrated into Swedish society. Previous research shows that some of the obstacles for the integration of immigrant women can be that they give birthto many children, are unskilled or have a low degree of education and that they tend to be living under patriarchalgender patterns. Another problem seem to be that some women are not even known as members of the municipalities.All of these problems are as well what was shown in my own study. I have used theoretical perspectives from Bourdieu, Elias & Scotson, Giddens, Roman and al-Baldawi in my analysis.Bourdieu have interesting thoughts about capital, habitus and field, which can help us to understand how individuals arebeing shaped and are given different opportunities to act in a special way or direction. Elias & Scotson describesthrough their study around established and outsiders how the process of integration can take place and what effects thatcan be shown for the opportunities to succeed in that part. Giddens, Roman and al-Baldawi then give us different waysto look at the patriarchy and family structures around the world. The result of this study shows that the willingness to integrate and be able to take care of your own support for living isan important part for the women for succeeding. For the other women, that don’t succeed, it turns out to be just like theprevious research has been shown. Gender patterns, many children and a low or no education skill all seems to be partof the issue. It is also suggested that the generosity of the Swedish welfare system might hinder rather than help someimmigrant women to become integrated into Swedish society.

Likelihood prediction for generalized linear mixed models under covariate uncertainty

This paper presents the techniques of likelihood prediction for the generalized linear mixed models. Methods of likelihood prediction is explained through a series of examples; from a classical one to more complicated ones. The examples show, in simple cases, that the likelihood prediction (LP) coincides with already known best frequentist practice such as the best linear unbiased predictor. The paper outlines a way to deal with the covariate uncertainty while producing predictive inference. Using a Poisson error-in-variable generalized linear model, it has been shown that in complicated cases LP produces better results than already know methods.