Due to the revisions of the guidelines of the course of study to improve communicative competence in foreign language learning in junior and senior high schools, development of listening comprehension tests focusing on oral communication abilities is needed. In 1995, we developed two trial listening comprehension tests focusing on oral communication and collected the test data from 758 high school students. After the analysis of the trial tests, we made some adjustments and designed a final test in which a marking-sheet system was adopted for grading. By comparing the results of our final test, which was administered to 800 high school students in 1996, with those of two trial tests, we can examine how the change of the testing conditions affected the test results of high school students. In this paper, the following conditions are discussed: 1) How often should the test takers listen to the monologues and dialogues on tape?-- once or twice?; 2) How should they answer the questions?-- in a written form or by a marking-sheet system?; 3) How do the changes in the dialogues affect the results of the test?, and 4) How do the adjustments of time for marking answers and reading directions affect the results of the test? According to our analysis, we find that the following conditions are important for designing a listening comprehension test: 1) It is desirable to have the students listen to the monologues and dialogues on tape only one time if they are provided with some visual aids. 2) There are no significant differences in the Discrimination Indexes between questions needing written answers and questions scored by a marking-sheet system. 3) A small change in the dialogue had a noticeable effect on the test results. Moreover, dialogues used must be realistic and interactive. 4) Adequate time for marking answers is needed most. This may suggest that Japanese high school students are weak in responding spontaneously.