Multiple imputation using Fully Conditional Specification (FCS) implemented by the MICE algorithm as described by Van Buuren and Groothuis-Oudshoorn (2011). The method provides built-in imputation models for continuous, binary, unordered categorical, and ordered categorical data, including two-level continuous data. Stef van Buuren authored the algorithm, which includes diagnostic plots to inspect imputation quality.
Use Cases
- Impute missing continuous variables using predictive mean matching or a normal model.
- Handle missing binary data using logistic regression models.
- Address missing unordered categorical data using polytomous logistic regression.
- Impute missing ordered categorical data using proportional odds models.
- Maintain consistency between imputed variables using passive imputation techniques.
Strengths
- Implements the MICE algorithm as described in a peer-reviewed 2011 journal article.
- Provides multiple built-in imputation models tailored to specific data types (continuous, binary, categorical).
- Includes diagnostic plots for inspecting the quality of imputations.
Limitations
- Description metadata is limited; actual data quality requires manual inspection after download.
- Column-level documentation is absent; field semantics must be inferred after download.
Provenance
- Source
- Stef van Buuren
- Collection Method
- Algorithm implementation and statistical methodology.
- Time Range
- null
- Freshness
- Last update date is unknown; freshness unverified.
- Geography
- null