Business Knowledge Series course

Presented by Patricia Berglund, senior research associate in the Survey Methodology Program at the Institute for Social Research at the University of Michigan

Concentrating on the needs of those relatively new to the use of multiple imputation tools in SAS, this course provides a general introduction to using the MI and MIANALYZE procedures for multiple imputation and subsequent analyses with imputed data sets.

Learn how to

• recognize the type of missing data patterns that exist in your data sets
• analyze imputed data sets using standard SAS procedures
• use PROC MIANALYZE to correctly analyze output from imputed files and subsequent procedure output from standard SAS procedures
• use real-world data sets in the virtual lab to obtain experience running SAS imputation procedures.

Who should attend

Analysts, data managers, and other data professionals working with data sets with missing data

Formats available	Standard duration
Classroom:	1.0 day
Live Web Classroom:	2 x 3 hour sessions	System Requirements

Before attending this course, you should

• have a solid understanding of the SAS DATA step, which can be gained by attending the SAS Programming 1: Essentials course or the SAS Programming 2: Data Manipulation Techniques course, or have equivalent experience
• have an intermediate knowledge of statistics, which can be gained by attending either the Statistics 1: Introduction to ANOVA, Regression, and Logistic Regression course or the Statistics 2: ANOVA and Regression course, or have equivalent academic training.

This course addresses SAS/STAT software.

Missing Data Issues

• types of missing data and how to identify sources and patterns of missing data
• why missing data occurs, and what to do about it

Introduction to Multiple Imputation Using SAS

• comparison of simple and multiple imputation approaches
• discussion of why multiple imputation is a preferred approach
• PROC MI and PROC MIANALYZE

Overview of Three-Step Process

• multiple imputation using PROC MI
• analysis of imputed data sets using standard SAS procedures
• use of PROC MIANALYZE for accounting for variability introduced during multiple imputation and analysis of output from standard SAS procedures

Practical Examples of Multiple Imputation

• common examples of multiple imputation and analysis of imputed data sets using public release data from the Longitudinal Survey of Aging (a complex sample survey data set)
• examples that cover typical imputation needs and subsequent analysis of imputed data using descriptive and regression approaches
• output from the imputation step and the analysis of imputed data sets