This paper examines the results of a questionnaire on reading ability in English by Japanese college students, which was formerly analyzed using raw scores, from the viewpoint of Rasch measured scores. In the Rasch analysis, the basic requirements for measuring are the following: (1) reduction of experience to one dimensional abstraction; (2) comparisons among persons and items; (3) the linear magnitude inherent in positioning objects along a line; and (4) a unit determined by a process that can be repeated without modification over the range of the variable. With these requirements in mind, this article deals with the rating scale data by focusing on the idea of linearity. Results are examined along a continuum. It is concluded that reading for testing has two components: contents and skills. Contents should be authentic materials because students use skills to comprehend or draw conclusions from the contents. The eventual scores given by the tests are the end product of their comprehension. Thus the process of reading is analyzed by examining three factors: students are given authentic materials; they struggle to comprehend the material and try to use it; their scores are calculated. (6 references.) (KFT)