How Digital Transformation Strengthens Regional Innovation Systems Through Public Research Platforms: Evidence from State Key Laboratories in China

Scientific problem. Although digital transformation is widely regarded as a system-level enabler of regional innovation, regions with similar digital investment still display markedly different innovation outcomes, and the organizational channels that convert digital conditions into innovation performance remain insufficiently specified. Aim and method. Using panel data for 30 Chinese provinces from 2014 to 2019, this study conceptualizes State Key Laboratories (SKLs) as a specific institutional form of public research platforms (PRPs) and examines whether SKLs transmit the effect of regional digital transformation to regional innovation performance. We construct a Regional Digital Transformation Index using the entropy weight TOPSIS method and estimate its effects using two-way fixed effects models, IV estimation, policy shock DID, mediation analysis, and fsQCA. Results. Regional digital transformation is positively associated with regional innovation performance, and DID estimates based on the National Big Data Comprehensive Pilot Zone policy point in the same direction, with the average policy effect significant at the 5% level under province-clustered inference (p < 0.05). Mediation analysis shows that SKLs transmit part of this effect through both boundary-spanning functions. The external knowledge introduction channel, captured by international collaboration, carries the larger indirect effect (318.966, Sobel z = 2.14, p = 0.032), while the internal network diffusion channel, captured by the network hub function, carries a smaller indirect effect (93.285, Sobel z = 1.89, p = 0.059) whose cluster bootstrap interval excludes zero. H3a (external collaboration) is confirmed by both the Sobel test and the BCa bootstrap interval, while H3b (network hub) is supported by the BCa interval and marginally by the Sobel test. fsQCA further shows that superior innovation performance arises from multiple configurations of digital transformation, SKL boundary-spanning functions, human capital, and industrial base. Conclusion and future research. The findings suggest that digital transformation acts less as a standalone technological input than as a system-level enabling condition whose innovation effects depend on the boundary-spanning functions of public research platforms and their alignment with regional conditions. Future research could use laboratory- and firm-level data and dynamic configurational methods to trace these mechanisms over time.