A distributed flat file strategy for managing research data
Ronald W. Helms · 1980
The purpose of this paper is to present a “Distributed Flat File” strategy and methods for designing effective systems for processing data from longitudinal research projects. First, let us consider the setting for which the strategy and corresponding techniques are intended.The methods described here are intended for longitudinal studies, those which involve studying the same subjects over a number of time periods, with data collected one or more times in each period. The length of the study is assumed to be such that irregularities will inevitably occur in the data collection: subjects will sometimes be unavailable for measurement; some subjects may perish or otherwise not complete the entire sequence of time periods, etc. It is also assumed that substantial deviations from the planned schedule may occur, and that some measurements may be made at random times, as for example, when an illness occurs.