The passage comprehensibility of three diagnostic reading tests--the Stanford Diagnostic Reading Test, the Spache Diagnostic Reading Scales, and the Durrell Analysis of Reading Difficulty--was evaluated using the following text variables: content familiarity, vocabulary, the number of idea units in the text, the number of text-based inferences required to integrate ideas, the existence of main idea statements, and the logical sequence of ideas in the passage. Selections from the second, fourth and sixth reader levels were evaluated to identify passage characteristics across reader levels and test inconsistencies. With regard to text variables, each selection was transformed into a series of propositions or ideas. Also, eight graduate students rated each selection for content familiarity. With regard to readability formula variables, a computer program was used in which each passage was evaluated in terms of four readability formulas: The ARI, Flesch-Kincaid, Harris-Jacobson, and the Dale-Chall. Results showed that the three tests differed substantially in the number of ideas presented, the number of inferences necessary to integrate these ideas, and text organization features. The readability indices reflected only surface features of the text and did not reflect the meaning characteristics of the passage nor identify the variables integral for deriving meaning from the text. (HOD)