Tests of reading ability have been devised--many of them. Their reliability coefficients have been determined and norms have been obtained; but do their reliability coefficients or their norms guarantee that they are good tests of reading ability or tests of reading ability at all? A reliability coefficient is only one criterion of a test. It measures the amount of agreement one would expect between an individual's score on one form of a test and his score on any other comparable form of the same test; but is there anything to show that the so-called reading tests do measure reading ability? The present studies were undertaken to get at a method by which we could determine whether the so-called reading tests do measure reading ability. (PsycINFO Database Record (c) 2016 APA, all rights reserved)