An investigation of reading comprehension tests was undertaken in this doctoral dissertation study to find out what the tests evaluate. The history of standardized reading tests and of efforts to define comprehension was traced. Standardized tests from three batteries (California Achievement Test, Gates-MacGinitie Reading Test, and Stanford Achievement Test) were analyzed at three grade levels (1 to 2, 4 to 6, 9 to 14). The first analysis, for readability, used the Dale-Chall and the Spache formulae, and the second analysis, for difficulty of test items, used specially designed rating scales. Tests differed on all readability counts and scores increased with higher grade levels, although not always at the same rates either across tests or within them. Test items tended to use different types of questions for similar kinds of materials. Questions fell generally into paraphrase and concept types. Frequency data were calculated for each of the readability and difficulty measures, and correlation coefficients were calculated for several characteristics of individual tests. While differences were found among the tests, they appeared to be testing abilities similar to those evaluated by intelligence and achievement tests. A model for constructing new and better defined reading comprehension tests, a bibliography, and tables are included. (MS)