Harry Ganzeboom will give a presentation on missing values:
Missing values: what is there beyond pairwise and listwise deletion?
“I discuss and compare several modes of missing values treatment in a model that predicts educational and occupation attainment in Suriname in which both mother’s and father’s characteristics (their education and occupation) are incompletely observed, for different reasons”
– Incomplete families when respondent was growing up (several types)
– Don’t know and refusal.
We compare several types of missing values treatment:
– Listwise deletion / complete case analysis
– Pairwise deletion (SPSS) / available case analysis
– Maximum likelihood estimation: FIML (Lisrel) and MLMV (Stata)
– Mean substitution
– Hot deck substitution
– Multiple imputation (SPSS)
Allison, Paul (2002)., Missing Data. Sage University Paper.
Enders, G.K. (2010). Applied Missing Data Analysis.
Knight, P. et al. (2007), Missing Values. A Gentle Intrdoduction”