ABSTRACT
In many disciplines, missing data are relatively uncommon and often taken as an indicator of sloppy science or poor methodology, and as a result techniques for dealing with missing data when they occur are generally frowned upon in these areas. In a conversation about dealing with missing data in a large interdisciplinary project, one of our colleagues, a geneticist, sarcastically remarked “All of our subjects had DNA.” Less than 2 weeks later, however, many of the laboratories at his institution were under several feet of water as a result of a hurricane. Fortunately, we had backup samples in multiple sites, his laboratory was on a higher floor of the building, and liquid nitrogen backup systems that required no external power source kept samples frozen until they could be rescued. Subsequent discussions regarding treatment of missing data have been better received.
