This study examined the effects of specially designed thinking journal activities that have been attributed with encouraging reflective thinking, on instruction using generic, or content-free problem solving software. Sixty-three fourth grade students participated in four instructional sessions using a software package called Moptown Hotel. Students completed separate posttests that measured (1) performance on problems of the same kind as those used in instruction, and (2) transfer of skills to other kinds of problems. Scores of students who wrote thinking journals prior to testing were compared with scores of students who did not. Results indicate that students who wrote thinking journals performed the same as students who did not when tested on problems similar to those practiced in class. Tests in which students transferred their skills to word problems, however, produced significant differences. There was no significant difference between scores when averaged over all four weekly occasions. However, for the final session alone, students who wrote thinking journals scored higher on tests of problem solving transfer than students who did not (p < .01). The study also examined the relationship between the degree of metacognitive thought displayed in students' journal entries, and their measured problem solving ability. Results indicate that students who had higher average reflectivity scores also had higher average problem solving performance and transfer scores (p < .05). It was also noted that the significant relationship between reflectivity and scores of problem solving ability was only observed in male students. It was concluded that under the right conditions, and for the right kinds of problems, thinking journal writing can help students understand their own thinking processes, resulting in improved problem solving behavior. The study also raises the question of whether there are differences between the ways that male and female students apply metacognitive awareness gained through journal writing experiences.