Page 1 of 1

Missing data: new Stat-JR functionality to support analyses of incomplete datasets

Posted: Thu Nov 24, 2016 4:05 pm
by CMM
There are now three principal Stat-JR templates available to support handling missing data in multilevel generalised linear models. These are typically much quicker than the equivalent executions in REALCOM-IMPUTE, and allow for greater flexibility too. Please note, though, that these templates have not been as widely-tested as REALCOM-IMPUTE.

The templates use two different approaches. The first two templates use ‘multiple imputation’ which is a widely used procedure that will handle a large number of models: a 2-level (2LevelImpute; available since 2014) and a new N-level (NLevelImpute) version are now available. The second (one pass) approach is a more recent generalisation with a more robust theoretical justification (see Goldstein et al, 2014 for further details); this has been implemented in the new 2LevelMissingOnePass template.

To use these templates, you will first need to install Stat-JR (http://www.bristol.ac.uk/cmm/software/s ... er-statjr/), and then download the zipped file containing the templates from CMM's missing data page (http://www.bristol.ac.uk/cmm/research/missing-data/).

Goldstein, H., Carpenter, J. R. and Browne, W. J. (2014), Fitting multilevel multivariate models with missing data in responses and covariates that may include interactions and non-linear terms. Journal of the Royal Statistical Society: Series A (Statistics in Society). 177(2), 553-564 http://dx.doi.org/10.1111/rssa.12022