Although each statistical unit on which measurements are taken is unique typically there is not enough information available to account totally for its uniqueness. Therefore heterogeneity among units has to be limited by structural assumptions. One classical approach is to use random effects models, which assume that heterogeneity can be described by distributional assumptions. However, inferen...

Meta-analysis is a standard statistical method used to combine the conclusions of individual studies that are related and the results of single study alone can not answered to deal with issues. The data are summarized by one or more outcome measure estimates along with their standard errors. The multivariate model and the variations between studies are not considered in most articles. Here we d...

This paper sheds some light on the problems associated with random effects gravity approaches. Arguments for the superiority of a fixed effects model are given both along intuitive and econometric lines based on a Hausman test.  2000 Elsevier Science S.A. All rights reserved.

This paper uses data drawn from the English Football League to model separate hazard rates for club managers for the 2001/2, 2002/3 and 2003/4 seasons. On average over the three seasons, approximately one-third of managers involuntarily exited employment status with their club. We model the hazard using a standard logistic model exploiting information on the spell at risk rather than the indivi...