Analysis of binary data for the balanced one-way classification |
| |
Authors: | Shah Arvind K PL Claypool |
| |
Institution: | 1. Department of Math &2. Stat , University of South Alabama , Mobile, Alabama, 36688 |
| |
Abstract: | A test is proposed for testing the equality of proportions based on the data available from a one-way classification having t treatment conditions and n binary observations per treatment. The test statistic B is a constant multiple of the F-statistic which results when the analysis of variance procedure for the one-way classification is applied to the data and, hence, is computationally simple. The statistic B from this binary analysis of variance (BIANOVA) is distributed asymptotically as a chi-square random variable. The proposed test is uniformly more powerful than either the F-test indicated above or the Pearson chi-square test; however, the attained empirical level of significance is frequently higher than for either of these competitors and usually higher than the stated level of significance for smaller values of n (say n ≤ 20). |
| |
Keywords: | equality of proportions chi-square test bianova one-way binary analysis |
|
|