TL20: Missing Values: Is there a difference between "Do Not Know" or "Choose not to Answer" and Responses left Missing and Should these responses be treated Differently?
*Tammy Massie, National Institute of Health 

Keywords: Missing Values, Imputation, Survey,

Missing values are common in surveys and questionnaires in a variety of settings and types of studies. Addressing and analyzing data that is missing is a challenge with many different and diverse options to perform sensitivity analysis including a variety of imputation methods. But in many surveys and questionnaires there are responses that include “Do Not Know” or “Choose Not to Answer”. Are these responses fundamentally different than missing values? Should these responses be treated and analyzed differently that missing values? Within this roundtable discussion the discussant leader will facilitate a discussion on what practices are commonly used and will attempt to brainstorm on what other options may be useful to consider.