A0565
Title: Sensitivity analysis of the choice of multiple imputation approach on categorical GPAbin biplots
Authors: Johane Nienkemper-Swanepoel - Stellenbosch University (South Africa) [presenting]
Niel Le Roux - Stellenbosch University (South Africa)
Sugnet Lubbe - Stellenbosch University (South Africa)
Abstract: Multiple imputation is generally considered as the recommended approach for the handling of missing data. This approach entails the computation of multiple completed data sets which are then analyzed separately by means of standard complete data techniques. Estimates from the separate analyses are combined using suitable combining rules. A popular method for combining the estimates is the so-called Rubin's rules. In the context of exploratory analysis, GPAbin biplots enable the combined visualization of multivariate visualizations constructed from multiple imputed data sets. This visualization approach combines configurations by means of generalized orthogonal Procrustes analysis (GPA) and applying Rubin's rules (-bin) to the aligned configurations. The performance of the GPAbin biplots has been evaluated using multiple imputation with multiple correspondence analysis (MIMCA) in an extensive simulation study. The performance of GPAbin will be evaluated when applying various multiple imputation techniques available in R. The effect of the choice of the multiple imputation approach will be investigated and presented by means of simulated examples.
