Randomization Procedures Applied to Analysis of Ballistic Data
Final rept. Jun 1990-May 1991
ARMY BALLISTIC RESEARCH LAB ABERDEEN PROVING GROUND MD
Pagination or Media Count:
Randomization procedures are shown to offer a viable approach to the analysis of ballistic data in a variety of settings. Distribution assumptions are unnecessary and, of even greater importance, random samples of data are not required. Small sample sizes, while never welcome, can also be accommodated. The price paid for randomization is increased computation, since every problem requires a tailored solution. Application of normal theory may only be valid to the extent that it approximates the results obtained from randomization. Reconciliation of some theoretical questions raised by the application of randomization procedures to ballistics data establishes this is an important and practical area of research.
- Statistics and Probability