Related Experiment Videos
Multilevel analysis with messy data
1De Montfort University, James Went Building 2-8, The Gateway, Leicester LE2 5YL, UK.
Statistical Methods in Medical Research
|January 5, 2002
Summary
This study reviews multiple imputation for handling imperfect multilevel data, including missing values and imprecise measurements. The method effectively uses all available information for robust statistical inferences.
Area of Science:
- Statistics
- Data Analysis
- Multilevel Modeling
Background:
- Multilevel data often contain imperfections like missing values and imprecise measurements.
- Standard statistical methods may not adequately address these data imperfections.
Purpose of the Study:
- To review the applications of multiple imputation for handling imperfections in multilevel data.
- To emphasize the role of data imperfection models in statistical analysis.
Main Methods:
- Review of multiple imputation techniques.
- Application to multilevel data with missing values.
- Application to multilevel data with imprecise measurements (corrupted recording).
Main Results:
- Multiple imputation can account for data imperfections and the processes underlying them.
- Inferences drawn from multiple imputation exploit all collected information.
Conclusions:
- Multiple imputation is a flexible method for dealing with imperfect multilevel data.
- This approach ensures that inferences appropriately reflect the information contained within the data.