Using partitioned databases for statistical data analysis

  • Authors:

  • Ruven Brooks;Meera Blattner;Zdzislaw Pawlak;Eamon Barrett

  • Affiliations:

  • University of Texas Medical Branch, Galveston, Texas;University of California, Davis, Livermore, California;Polish Academy of Sciences, Warsaw, Poland;Jaycor Corporation, Washington, D.C.

  • Venue:
  • AFIPS '81 Proceedings of the May 4-7, 1981, national computer conference
  • Year:
  • 1981

Quantified Score

Hi-index 0.00

Visualization

Abstract

The statistical analysis of scientific data is a process that can be viewed as consisting of three fundamental phases. First, the observations are recorded. Next, they are encoded into a numeric form suitable for statistical analysis. Finally, the calculations are performed for the particular type of analysis needed for the design of the study. This ordering is, however, only conceptual; in most real studies, the three phases interact and are overlapped. Thus, it may be the case that a preliminary analysis run indicates that a more refined coding scheme is needed or that the coding process reveals deficiencies in the data collection.