A generalized analysis of variance program utilizing binary logic

  • Authors:
  • P. Real

  • Affiliations:
  • Westinghouse Electric Corporation, Baltimore, Maryland

  • Venue:
  • ACM '59 Preprints of papers presented at the 14th national meeting of the Association for Computing Machinery
  • Year:
  • 1959

Quantified Score

Hi-index 0.00

Visualization

Abstract

Analysis of variance is a statistical technique for determining the significant sources of variation of experimental test results where each test result is dependent on a number of different test conditions. The computations required in performing such an analysis are simple and straightforward but extremely tedious when performed manually. For larger experiments manual computation is prohibitive. While the analysis for any single experiment may be easily programmed, the resulting program will be of little use for other experiments involving even slight variations.A completely generalized program is the subject of this paper. The first paragraphs review the conventional notations and formulas. These formulas are then generalized to obtain the unique computational method used for their evaluation.