1997 年 3 巻 p. 147-156
"Journal Writing" is one of the methods to use writing in mathematics classes, and it has two functions; "(1) Expressive function", and "(2) Transactional function". In this paper, "(2) Transactional function" is focused on, and the case described in Ninomiya (1996 (c)) is investigated. First of all, "the framework for the investigation of Transactional Writing", which has four stages; impression, summary(1), summary(2), consideration, is established from the framework of Nakamura (1989). Since the writing activities are taken place during the second term of 1996, the difference of the paper tests' results between the first and the second terms should be considered as to be caused by the writing activities. Every student is over the national average during the second term while more than one-fourth of the students were below national average during the first term. Furthermore, the students who are not good at math improves their score a lot, which means the advantages of writing activities work very effectively. It also means that writing activities makes "the education for every student" policy possible. Writings which have written by students are analysed in two ways; by the class teacher and from "the framework for the investigation of Transactional Writing". The correlation of these two analysis' results is 0.846. Since the analysis by class teacher is reliable, the reliability of the framework is suggested. The correlation between writing and mathematical ability is also high. This means that mathematical writing is defferent from writings in language classes. Moreover, mathematical writing ability is correlated with higher-order abilities such as an attitude or way of thinking, a lot. From such outcome, mathematical writing ability should be considered as one of the higher-order abilities, which is stressed in the current curriculum for school mathematics.