Longitudinal Data Modelling Using Penalized Splines And Ranked Set Sampling

Longitudinal Data Modelling Using Penalized Splines and Ranked Set Sampling

"Longitudinal studies, where data is collected by measuring the same experimental units several times over a relatively long period, are becoming increasingly common. Conventional statistical approaches have limitations when applied to the analysis of longitudinal data ... Practical limitations of longitudinal analysis that relate to missing data and large data set sizes were explored in this thesis with the application of a sampling technique known as Ranked Set Sampling (RSS). We developed this sampling method, which has not previously been applied to longitudinal data, for fixed and mixed-effects models. This thesis also illustrated inference techniques to estimate these models after selecting sample units by RSS." --Abstract.

The UNICEF-WHO-World Bank Joint Child Malnutrition Estimates (JME) standard methodology

This document provides the background, rationale and description of the standard approach followed by the UNICEF-WHO-World Bank Joint Malnutrition Estimates (JME) group to generate national estimates for Sustainable Development Goal (SDG) indicators 2.2.1 (child stunting), 2.2.2 (1) (child overweight) and 2.2.2 (2) (child wasting), as well as regional and global aggregations for the three indicators. The JME process for generating national, regional and global estimates is described along with: - compilation of data sources with anthropometric data - production of estimates of child malnutrition and data quality measures through use of standardized methods applied to country microdata when available - review of anthropometric data sources, considering data collection methodology and data quality assessment and trends inclusion criteria for data sources - the child malnutrition database - modelled estimates for child stunting and child overweight - production of national, regional and global trends

Biometrics