Katz and Allen (1982) found that the performance of R & D projects had peaked at 2 to 4 years of mean tenure of project team members, and declined significantly after the years. They considered that the phenomenon occurred because of the tendency of a project group of stable composition to believe it possesses a monopoly of knowledge of its fields, and named it "Not-Invented-Here (NIH) syndrome." As compared to the widely spread of this concept, significant faults existed in their empirical method has not adequately been recognized. In fact, we do not yet have absolute proofs of whether the NIH syndrome really exists or not. Confronting the difficulty of the method which Katz and Allen had taken, I put a high value on large quantity of patent data, and analyze it to reinterpret the NIH syndrome from an individual perspective. First, I create the concept of "co-inventive duration," which designates the average duration of maintaining relationships with the focal person's co-inventors, and use it to see if Katz and Allen (1982)'s empirical result (KA-test) can be confirmed. The results showed that inventors' performance on quantity basis peaked at 2 to 4 years of the co-inventive duration, which supported the KA-test, however, qualitative performance improved linearly with the co-inventive duration, which showed no support for the KA-test. Next, I took a field study to investigate backgrounds of the former results at a Japanese machinery firm. Inventors who worked for the firm were asked to rate their perceived existence of the NIH component and team building component-both were considered to be the causes of the KA-test. But correlation with the co-inventive duration did not reveal strong supports for these components. Then, I directly asked inventors who had relatively longer co-inventive duration about detailed relationship with each of their co-inventors, and found that they maintained relationships to fuse technical and customer information together. From these results, I concluded that the NIH syndrome might not exist in terms of qualitative performance.