Please join us for our monthly Biomedical Data Science Grand Rounds with Jerry Reiter, PhD, Department Chair and Professor of Statistical Science, Duke University on Thursday, November 21 at 12:00pm. 

Talk title: “How Auxiliary Information Can Help Your Missing Data Problem”

Location: Williamson 571 E/W, DHMC (exit Williamson Level 5 elevators, turn left, and walk down the hallway, past the kitchen)

A light lunch will be provided on a first-come, first-served basis. 

Talk Summary 

Many surveys (and other types of databases) suffer from unit and item nonresponse. Typical practice accounts for unit nonresponse by inflating respondents’ survey weights, and accounts for item nonresponse using some form of imputation.  Most methods implicitly treat both sources of nonresponse as missing at random.  Sometimes, however, one knows information about the marginal distributions of some of the variables subject to missingness.  In this talk, I discuss how such information can be leveraged to handle nonignorable missing data, including allowing different mechanisms for unit and item nonresponse (e.g., nonignorable unit nonresponse and ignorable item nonresponse).

Biography

Jerry Reiter is Department Chair and Professor of Statistical Science at Duke University. His main areas of research include methods for protecting confidentiality in data, for handling missing and erroneous values, for integrating information from multiple sources, and for analysis of surveys and causal studies.  He received a PhD in statistics from Harvard University in 1999.