Athens University Economics and Business (AUEB)
Goodness of fit tests for random effect models with binary responses estimated via h-likelihood
Abstract
dc:descriptionThe objective of this study is to introduce goodness of fit statistics for the systematic part of random effect models. These statistics are formed as score tests and are based on partitioning the observations into mutually exclusive groups. Weighted versions of these statistics are also introduced which are based on the correlation between an appropriately adjusted candidate covariate for entrance into the model and the model residuals. In addition to this, weighted cumulative and moving sum processes are also provided with a goal to detect, in a more effecient way, various departures from the null model, as compared to the usual cumulative and moving sums. The estimating procedure that is used throughout this study is the h-likelihood, so the quantities needed for the above mentioned statistics are derived based on this estimating approach. The simulations were designed so as to examine the performance of the tests under different factors, such as sample size, magnitude of random effects' variance, type of the model misspecification, etc. The results indicate the superior performance of the weighted tests of any type (either those based on partitioning or those based on cumulative sums), there are cases, however, were some unweighted tests displayed a similar power as their weighted analogous. The proposed tests are illustrated with the use of two real datasets.
Degree
thesis:*- Grantor dc:publisher
- Athens University Economics and Business (AUEB)
- Year dc:date
- 2017
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Korre, Antonia
- Κορρέ, Αντωνία
Subjects
dc:subject × 6Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/40322