The reliability of a test is usually defined as the consistency with which a test measures whatever it measures, the consistency being indicated by the correlation between two administrations of the same test or between equivalent forms. Whenever it has not been feasible to administer the test twice, or whenever full-length equivalent forms have not been available, the procedure usually followed has been that of making two synthetic tests by scoring the odd- and even-numbered items separately, or by scoring two randomly-drawn lists of items. Dividing the test in this manner results in obtaining a correlation between two forms which are half the length of the original test. It has long been recognized that the reliability of a test is increased as its length is increased. Spearman (17) and Brown (1), in a formula developed independently, have expressed the degree of reliability to be expected by lengthening tests. When this formula is applied to the correlation obtained from two halves of a test, an estimation of the reliability to be expected by lengthening the tests is obtained.
Proceedings of the Iowa Academy of Science
©1949 Iowa Academy of Science, Inc.
Neidt, Charles O.
"Homogeneity of Behavior Indicated from the Spearman-Brown Prophecy Formula,"
Proceedings of the Iowa Academy of Science, 56(1), 271-277.
Available at: https://scholarworks.uni.edu/pias/vol56/iss1/38