A Validation Test of the Uzzi Et Al. Novelty Measure of Innovation and Applications to Collaboration Patterns Between Institutions

Yuefen Wang,Lipeng Fan,Lei Wu

DOI: https://doi.org/10.1007/s11192-024-05071-7

IF: 3.801

2024-01-01

Scientometrics

Abstract:Exploring a robust and universal appeal bibliometric indicator for assessing creativity is essential but challenging. The novelty measure of innovation proposed by Uzzi et al. (NoveltyU) has sparked considerable interest and debate. Thus, further validation and understanding of its portfolio form of novelty and scope of application are necessary. This paper delves into the calculation and application of the NoveltyU method to shed light on its effectiveness and scope. Analysis of the calculation process reveals that journal pairs with higher novelty often span independent fundamental areas, while those with lower novelty tend to focus on similar and applied fields. Utilizing collaboration patterns between institutions, as discussed in our prior study (Fan et al., Scientometrics 125:1179–1196, 2020), offers insight into the method’s performance in real-world contexts. Results consistently show higher mean NoveltyU values in MM pattern over time, affirming the method’s validity. Categorizing papers into high conventional, low conventional, low novel, and high novel categories unveils higher overlap degree of terms among different patterns in high novel papers. Moreover, leading terms in MM pattern exhibit specific information, while delay terms tend to be more general, and simultaneous terms are even more so. These findings offer valuable insights into identifying hot and frontier topics, bolstering the credibility and application potential of the NoveltyU method, aligning with the broader objective of establishing valid measures of innovativeness in research.