ESTIMATION OF PARAMETERS IN MULTIVARIATE EXPERIMENTAL DESIGN AND A HYPOTHETICAL APPLICATION TO ESTIMATING FRACTIONAL DOSE RECOVERY EFFECTS
Abstract
A statistical technique is explained, derived and illustrated. It is believed that a study of the values of the terms of the assumed mathematical model (or the graphs therefrom) would be of great informational help to the scientist in interpreting his results. Analysis of variance, components of variance analysis, and the estimation technique described herein applied to a properly designed experiment will go far toward extracting all the relevant information from a sample. Estimation formulas for the terms of the assumed mathematical model are derived for two of the most used statistical experimental designs, the factorial and the orthogonal squares. The method is illustrated on a three-variable, hypothetical dose recovery experiment. It is pointed out that the method gives one a handle on measuring quantitatively the separate and joint contributions to the observed effect, of variables otherwise hopelessly correlated and overlapping.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1966
- Accession Number
- AD0635167
Entities
People
- J. A. Greenwood
Organizations
- Armed Forces Radiobiology Research Institute