University of Twente Student Theses

Login

A comparison of multivariate and univariate models for pre-test post-test data concerning accuracy in the presence of missing data

Beinhauer, L.J. (2018) A comparison of multivariate and univariate models for pre-test post-test data concerning accuracy in the presence of missing data.

[img] PDF
322kB
Abstract:Missing data has been posing a common problem to almost all scientific studies over the last decades. This study compares multivariate models to univariate models, the change score method and the regressor variable method, in terms of precision and accuracy over different conditions of missing data. In two simulation studies patterns of missing data were introduced, over degrees of missingness and covariance for pre-test post-test designs. The first study focuses on the MCAR pattern and its influence on the estimates of the different models. The second model makes use of a MAR pattern, where the likelihood of missing values is dependent on an observed variable. The results showed some differences between the models. The change score method appeared to have no particular advantages, the regressor variable method had the highest accuracy, while the multivariate method provided the best precision. Higher amounts of missing data had a negative impact on both accuracy and precision.
Item Type:Essay (Bachelor)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:77 psychology
Programme:Psychology BSc (56604)
Link to this item:https://purl.utwente.nl/essays/75269
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page