We analyze the mechanism in which the productivity of a researcher or a research group is determined. For that purpose, we made a quantitative comparison between two research groups in the field of post-world-war theoretical astrophysics in Japan.
Our main conclusion is summarized as follows. The reputation of a research group within the research community (in our case, the community of Japanese astronomers) does not necessarily coincides with the macroscopic indices such as the number of publications or the average citation index. In other words, the two groups showed similar scores for these macroscopic indices. The number of highly-cited papers shows some difference, but not as large as the difference in the reputation. This result indicates that the quantitative measurement of the productivity does not give meaningful measure for the actual contribution of a research group to science.
In order to investigate the difference between two groups, we compared the ``field coverage'' of the two groups. In both groups, the field coverage of a person and his/her productivity (in number of publications) show strong positive correlation. Moreover, there is big difference in the average coverage between two groups. In our particular example, the field coverage of a group is related to the number of new research field pioneered by the group. This implies that the reputation of a group within the research community depends more on the number and the wideness of the new field pioneered by the group.