A Sequential Method for Screening Experimental Variables
- 1 June 1962
- journal article
- research article
- Published by JSTOR in Journal of the American Statistical Association
- Vol. 57 (298) , 455
- https://doi.org/10.2307/2281652
Abstract
This paper describes a sequential method which may be used to determine with efficiency the important variables and interactions in an experiment. The method is useful in certain cases involving up to 100,000 or more variables. Mathematical derivations are given and tables are included showing the optimum designs and the total numbers of tests or samples required. A simple case is first given, followed by cases with interactions or high-level effects.Keywords
This publication has 0 references indexed in Scilit: