The Fundamentals of Dealing With Missing Data
Missing data can lead to biased results if not properly considered. This is particularly important given that most standard analyses — e.g., analysis of variance (ANOVA) and regression — assume that missing data is not an issue. Researchers may find significant results that are not true or miss associations in the population. More problematic is that until researchers use modern missing data techniques consistently, we do not have a way to know how the bias will impact results.
During the first part of this webinar, presenter Isaac Washburn, Ph.D., will focus on the dangers that come from ignoring missing data issues. He’ll introduce the concepts of missing completely at random (MCAR), missing at random (MAR), and missing not at random (MNAR) and he’ll discuss how past methods of dealing with missing data (e.g., listwise deletion and mean substitution) are inadequate for dealing with the bias created by most forms of missing data.
The second part of the webinar will focus on the two modern methods of missing data: multiple imputation and maximum likelihood. Dr. Washburn will first focus on multiple imputation, as the principles of missingness are more clearly articulated and utilized with multiple imputation. Finally, Dr. Washburn will walk through several examples using Mplus, Stata, and SPSS, showing how multiple imputation works and highlighting maximum likelihood techniques.
Approved for 1.5 CFLE contact hours of continuing education credit.
By attending this webinar, you will:
- Understand the dangers of ignoring missing data
- Understand the principles of using multiple imputation
- Understand the principles of using maximum likelihood
- Family Scientists looking to increase their ability to deal with missing data
- Students or professionals learning about statistics
- Professors looking for a classroom resource for teaching statistics
About the Presenter
Isaac Washburn, Ph.D., is an assistant professor and research methodologist at the Department of Human Development and Family Science at Oklahoma State University. He has taught advanced research methods (e.g., measurement construction and study design) and statistics (e.g., Multilevel Modeling and Structural Equation Modeling) to doctoral students for the past nine years. He is also co-director of the Research Design and Analysis Core of the NIH-funded Center for Integrative Research on Childhood Adversity, a research-focused center headquartered in Tulsa, Oklahoma. Dr. Washburn received his Ph.D. at Oregon State University in human development and Family Science with an emphasis on advanced statistical modeling, and his B.S. in economics from Brigham Young University.
On-Demand Webinar Recording and Classroom Use
Even if you can't watch this webinar live, your registration will still grant you access to watch the recording at your convenience. The fee for this webinar is $25 for NCFR student members, $45 for NCFR members, $85 for nonmembers.
License for classroom use by one professor is available for $105 for NCFR members, $185 for nonmembers.
License for departmental use (multiple professors) is available for $155 NCFR member / $305 nonmember.
Departmental license for CFLE-approved programs is $125.