The aim of this paper is to present linguists--both members of research teams and language teachers--with some possible applications of computing science to language teaching. First, it is possible to use a computer's output system merely in order to print any type of information, e.g., a list of students' names and their marks. The computer also allows the teacher to make more elaborate calculations and obtain very useful results, such as the arithmetic mean, the standard deviation and so on, whatever the number of students. To illustrate clearly these applications, a concrete example is given concerning the improvement of a reading comprehension test. The test is fully described and its evaluation over three years is presented and explained: modifications of all sorts were introduced in accordance with the interpretation of mathematical and statistical data. The discrimination factor makes it clear whether any particular item discriminates effectively; the frequency factor shows the number of answers--right or wrong--given for each exercise. Finally, the correlation factor shows whether all the different series of exercises really contribute to testing the same kind of reading skill. All these data are plotted or given in charts, and are thoroughly discussed in order to explain how it is possible to transform such skill-testing material to make it as concise as possible while giving it as great a discriminatory power as possible. (Author/CLK)