This paper examines existing research regarding the problem of school dropouts. It examines the historical context of the problem and discusses current definitions and methods for identifying dropout characteristics and computing dropout rates. Few incentives exist for accurate record-keeping on dropouts. School finances are often linked to attendance reports, a circumstance that creates a conflict of interest for schools and administrators. Also, the varied methods districts use to determine enrollment alter the reported dropout rates. At the state level, education agencies should consider establishing common dropout definitions, methods for collecting data, and student tracking systems. In Maine, officials have become increasingly active in attempts to lower the dropout rate, which is reported annually as a percentage of fall enrollment. The state's average rate has stabilized at about 4% during the past 5 years. National computations show Maine's 79% graduation rate as being relatively high compared to the 71% national average. Many researchers have attempted to identify national dropout characteristics, looking at such factors as academic achievement, socioeconomic background, family background, individual attitudes, and ethnicity. Responsibility for the problem has been placed alternatively on the schools, society, and students themselves. An analysis of dropout programs and literature reveals a combination of prevention strategies, including non-punitive approaches, alternative school schedules, and modifying or rescinding policies that tend to discourage at-risk students. The traditional dropout prevention method has been segregation and remediation. The document concludes by calling for consensus on a national dropout definition, systematic research, and better understanding of the problem. It includes a bibliography with approximately 100 citations. (TES)