A text-based monitoring tool for the legitimacy and guidance of technological innovation systems

Abstract

Disruptive technologies have to overcome their liability of newness and transition into their growth phase by achieving compliance with existing institutions and pursuing the most promising development paths. Technological innovation system (TIS) studies examined these two issues of legitimacy and guidance of innovation activities by investigating the public discourse with manual media analyses. However, these approaches are time-consuming, prone to subjectivity biases and limited in scope. Therefore, our paper proposes an automatic text analysis methodology based on unsupervised Latent Dirichlet Allocation (LDA) topic modelling and lexicon-based sentiment analysis. By processing 3423 German newspaper articles from the Nexis Uni database, we cover the development of battery-electric vehicles (BEV) in Germany from 2009 to 2019 and identify five socio-technical aspects. Our results indicate an intact legitimacy for the TIS, with Usability, R&D, and Industry being legitimate aspects, which also exhibit strong or improving guidance. In contrast, the Infrastructure and Policy aspects have been less legitimate and weak in guidance, suggesting the need for more holistic policy measures and infrastructure expansion to establish a mass market. Our proposed methodology adds to the toolbox of methods to analyze TIS and serves as a monitoring tool to reveal contested aspects and periods in the public discourse.