This article on Post-Traumatic Stress Disorder (PTSD) discusses dealing with incomplete research data. From the article: "A common dilemma in many types of research comes about from the need to deal with incomplete or missing data. Within a single data collection session, missing data often comes in the form of a participant's refusal to answer a specific interview question, skipped items or scales (either intentionally or unintentionally) on a paper-and-pencil survey, or perhaps an early termination of the session, with only partially completed tasks or instruments. In the case of longitudinal research with multiple data collection sessions, missing data may be more pervasive, with respondents providing data sporadically across time or dropping out of the study altogether. It might even be that some participants enter a longitudinal study late, such that their data are not available for the initial stages of the research. Traditionally, the problem of missing data has been seen as a costly nuisance and sometimes even a 'fatal flaw' to a research project. [...] Recent statistical work has allowed for an almost revolutionary shift in the manner in which missing data can be handled. More important, the execution of contemporary missing data techniques can obviate many practical and statistical concerns. Modern missing data strategies enhance efficiency, preserve resources, and guard against incorrect statistical inference. In fact, one could argue that the incorporation of corrective missing data analyses and even purposeful missing data designs shortly will be considered routine and expected by scientific editorial boards and funding review panels. In the sections that follow, we describe conditions under which data are missing and present some of the newer maximum-likelihood based methods. We then discuss the implementation of some of these methods in the context of stress and trauma research. We also provide resources for those wishing to learn more about the methods."
National Center for PTSD: http://www.ptsd.va.gov/
PTSD Research Quarterly (Spring 2001), v.12 no.2